MBL
CONTENTS
No. 1, FUBRI ARV 1992
BEHAVIOR
Hermans, Colin O., and Richard A. Satterlie
Fast-strike feeding hehav ior in a pteropocl mollusk.
<-/ini/i: IniHiiiiiii Phipps
Wayne. Nancy L., and Gene D. Block
Effects of photoperiod and temperature on egg-lav-
ing behavior in a marine mollusk. .\/il\siei californica
DEVELOPMENT AND REPRODUCTION
Amemiya, S., and R. B. Emlet
I lii' development and larval form of .111 ec liinoihu-
rioid c'c lunoid, Astlienosoma ijiinin, revisiti-d
Ausio, Juan
I'm ilic.ilion and biochemical characterization of ihe
mu lear sperm-specific proteins ol the- bivalve mol-
\\isk.sAgl~iodesma saxicola and .M\iiliiiii'ini inil/nt/i ....
Blades-Eckelbarger, Pamela I., and Nancy H. Marcus
The origin of conical vesicles and then role- in egg
envelope formation in the "spiny" eggs of a calanoid
copepod. Centropagei velificatus
Chandler. Resa M., Mary Beth Thomas, and Julian
P. S. Smith, III
The role of shell granules and accessory cells in
eggshell formation in (.'.<m\'»luta pulrhra (Turbellaria,
Ac oela)
Chia, Fu-Shiang, Ron Koss, Shauna Stevens, and Jeff
I. Goldberg
Isolation ol neurons of a nudibranch veliger ....
Holland, Linda Z., and Nicholas D. Holland
Early development in ihe lancelet (=amphioxus)
Branchiostoma /inri/lac from sperm eniiv iliiough
pronuc lear fusion: presence of vegetal pole plasm
and lac k ol conspicuous ooplasmic segregation . .
Lee, Youn-Ho, and Victor D. Vacquier
The divergence of species-specific abalone sperm
Ivsins is promoted bv positive Darwinian seleiiion
ECOLOGY AND EVOLUTION
Gil-Turnes, M. Sofia, and William Fenical
1-iubivos of HniiKinis aiiu'i'icaiiu* are proteiti-<l bv
epibiotic bacteria
31
41
(ili
Williams-Howze, Judy, and Bruce C. Coull
Are temperature and photoperiod necessary cues
for encystmi'iil in llic marine' bentlm harpacticoid
copepod Hctt'iiiji^'llii^ I/HUH/ Coull?
GENERAL BIOLOGY
Jennings, Joseph B., Lester R. G. Cannon, and
Adrian J. Hick
I lie nature and origin ol ihe epidei mal scales of
Notodactylus handschini—3.n unusual temnocephahd
turbellarian ectosymbiotic on crayfish from north-
ern Queensland
Mangum, Charlotte P., James M. Colacino, and
Judith P. Grassle
Red blood cell oxygen binding in lapitclhd poly-
chaetes .
PHYSIOLOGY
Singarajah, K. V., and F. I. Harosi
Visual cells and pigments in a demersal fish, the
blaik sea bass (O;//n</)jn//\ \lrinlu)
Tankersley, Richard A., and Ronald V. Dimock, Jr.
Quantitative anahsis ol (he- structure and function
of the marsupial gills of (he freshwater mussel An-
i tiliii'intii .
117
129
135
145
RESEARCH NOTES
Feldgarden, Michael, and Philip O. Yund
Allorecognition in colonial marine invertebrates:
does selection favor fusion wiih km. or fusion with
self?
Rands, M. L., A. E. Douglas, B. C. Loughman, and
R. G. Ratcliffe
Avoidance of hypoxia in a cnidarian symbiosis In
algal photosvnthclic o\\gen
1115 The Biological Bulletin Board
POETRY
Skinner, Dorothy M.. and John S. Cook
Carroll M. Williams
Mellon, Deforest, Jr.
How tin- axon got its tale
CONTENTS
No. 2, APRIL 1992
165
167
Van Alstyne, Kathryn L., Chad R. Wylie, Valerie J.
Paul, and Karen Meyer
Antipredator defenses in tropical Pacific soft corals
(Coelenterata: Alcyonacea). I. Sclerites as defenses
against generalist carnivorous fishes 231
DEVELOPMENT AND REPRODUCTION
Hand, Cadet, and Kevin R. Uhlinger
The culture, sexual and asexual reproduction, and
growth of the sea anemone Nematostella i'i'iieii\i\ 169
McEdward, Larry R.
Morphology and development of a unique type of
pelagic larva in the starfish P/ennti'i /r\v7<//f<\ (Echi-
nodermata: Asteroidea) 177
ECOLOGY AND EVOLUTION
Jeffries, William B., Harold K. Voris, and Sombat
Poovachiranon
Age of the mangrove crab Sc\lla <n'mitn at coloni-
sation by stalked barnacles of the genus ()itul(i\mi\ 188
Kim, Kiho, Walter M. Goldberg, and George T.
Taylor
Architectural and mechanical properties of the black
coral skeleton (Coelenterata: Antipatharia): a com-
parison of two species 195
Raimondi. Peter T.
Adult plasticity and rapid larval evolution in a re-
cently isolated barnacle population 210
Shapiro, Daniel F.
Intercolony coordination of /ooid behavior and a
new class of pore plates in a marine brvo/oaii ... 221
NEUROBIOLOGY AND BEHAVIOR
Diaz-Miranda, Lucy, David A. Price, Michael J.
Greenberg, Terry D. Lee, Karen E. Doble. and Jose
E. Garcia-Arraras
Characterisation of two novel neuropeptides from
the sea cucumber Holotlniritt gluht'i 'mini 241
Mackie. G. O., C. E. Mills, and C. L. Singla
Giant axons and escape swimming in Eujilnl;iiiiii\
dunlapae (Ctenophora: Cydippida) 248
Saigusa, Masayuki
Phase shift of a tidal rhythm by light-dark cycles in
the semi-terrestrial crab Si'f-nruui /tntin/i 257
PHYSIOLOGY
Baker, S. M., and R. Mann
Effects of hypoxia and anoxia on larval settlement,
juvenile growth, and juvenile survival of the oyster
Crossostrfd I'lt^nuitt
Brown, A. Christine, and Nora B. Terwilliger
Developmental changes in ionic and osmotic regu-
lation in the Dungeness crab. C.auicr magintcr .... 270
Cronin, Thomas W.
Visual rhythms in stomatopod crustaceans observed
in the pseudopupil 278
No. 3, JUNE 1992
DEVELOPMENT AND REPRODUCTION
Fong. Peter P., and John S. Pearse
Evidence for a programmed circannual life cycle
modulated by inc reasing da\ lengths in Xi'iinll/i^ Inn-
i!;ro/fl(Polychaeta:Nereidae) from central California
Mita, Masatoshi, and Masaru Nakamura
Ultrastru< tural study of an endogenous energy sub-
strate in spermatozoa of the sea urchin Hi'i>iicfiitn>tit\
juili In'ii nini\ 298
Rivcst, Brian R.
Studies on the- struc tine and function of the lar\al
kicliic-\ complex of prosobranch gastropods 305
MARINE CELL BIOLOGY
Gates, Ruth D., Garen Baghdasarian, and Leonard
Muscatine
Temperature stress causes host cell detachment in
symbiotic cnidarians: implications tor coral bleach-
ing 324
NEUROBIOLOGY AND BEHAVIOR
Mercier, A. Joffre, and Rune T. Russenes
Modulation of crayfish hearts 1>\ FMRFamide-
velated peptides 333
CONTENTS
Kulk.it m, Gunderao K., and Milton Fingerman
Quantitative analysis by reverse phase high perfor-
mance liquid chromatography of 5-hydroxytrypt-
, innne in the central nervous system of the red
swamp uavfish, Procambanti dnrkii 341
Page, Louise R.
New interpretation of a nudibranch central nervous
system based on ultrastructural analysis of neuro-
developinent in Mi-lih<- Ifminiti. I. Cerebral and vis-
ceral loop ganglia 348
Page, Louise R.
New interpretation of a nudibranch central nervous
system based on ultrastructural analysis of neuro-
developim-iit in Mtlibeleonma. II. Pedal, pleural, and
labial ganglia 366
PHYSIOLOGY
Bergles, Dwight, and Sidney Tamm
Control of cilia in the branchial basket of dona in-
tr\tiiniln (Ascidacea) 382
Latz, Michael I., and James F. Case
Slow photic and chemical induction of biokmiines-
cence in the midwater shrimp, Sergeste* unnlis Han-
sen " 391
Fitt, W. K., and S. L. Coon
Evidence for ammonia as a natural cue for recruit-
ment of oyster larvae to oyster beds in a Georgia
salt marsh 401
Burton, Ronald S.
Proline synthesis during osmotic stress in megalopa
stage larvae of the blue crab, Callinectes wpulm . . 409
Combs, Christian A... Nicole Alford, Angela Boynton,
Mark Dvornak, and Raymond P. Henry
Behavioral regulation of hemolvmph osmolarity
through selective drinking in land crabs, Birgus Intro
and Gecarcoidea lalanrlii 416
Ellers, Olaf, and Malcolm Telford
Causes and consequences of fluctuating coelomic
pressure in sea urchins 424
Kraus, David W., Jeannette E. Doeller, and Jonathan
B. Wittenberg
Hydrogen sulfide reduction of symbiont cytochrome
<'552 i" gills of .S'n/ciww ri'nli (Mollusca) 435
Wilmot, David B., and Russell D. Vetter
Oxygen- and nitrogen-dependent sulfur metabolism
in the thiotrophic clam Snlemw reidi 444
VIEWS AND DISCUSSION
Grosberg, Richard K.
To thine own self be true? An addendum to Feld-
garden and Yund's report on fusion and the evo-
lution of allorecognition in colonial marine inver-
tebrates 454
Yund, Philip O., and Michael Feldgarden
To thine own self be true? Yes! Thou canst not then
be false to any other. A reply to Grosberg 458
Index to Volume 182 . 460
Volume 182
THE
Number 1
BIOLOGICAL
BULLETIN
fvlarine Biological Lat/uratory
LIBRARY
MAR 3 irtZ
Woods Hole, Mass.
FEBRUARY, 1992
Published by the Marine Biological Laboratory
Marine Biological Laboratory
LIBRARY
MAR 3 1992
Woods Hole, Mass.
THE
BIOLOGICAL BULLETIN
PUBLISHED BY
THE MARINE BIOLOGICAL LABORATORY
Associate Editors
PETER A. V. ANDERSON, The Whitney Laboratory, University of Florida
DAVID EPEL, Hopkins Marine Station, Stanford University
J. MALCOLM SHICK, University of Maine, Orono
Editorial Board
DAPHNE GAIL FAUTIN, University of Kansas
WILLIAM F. GILLY, Hopkins Marine Station, Stanford
University
K. RANGA RAO, University of West Florida
STEVEN VOGEL, Duke University
Editor: MICHAEL J. GREENBERG. The Whitney Laboratory, University of Florida
Managing Editor: PAMELA L. CLAPP, Marine Biological Laboratory
FEBRUARY, 1992
Printed and Issued by
LANCASTER PRESS, Inc.
PRINCE & LEMON STS.
LANCASTER, PA
THE BIOLOGICAL BULLETIN
THE BIOLOGICAL BULLETIN is published six times a year by the Marine Biological Laboratory, MBL
Street, Woods Hole. Massachusetts 02543.
Subscriptions and similar matter should be addressed to Subscription Manager, THE BIOLOGICAL BUL-
LETIN. Marine Biological Laboratory, Woods Hole, Massachusetts 02543. Single numbers, $30.00. Sub-
scription per volume (three issues). $77.50 ($155.00 per year for six issues).
Communications relative to manuscripts should be sent to Michael J. Greenberg. Editor-in-Chief, or
Pamela L. Clapp, Managing Editor, at the Marine Biological Laboratory, Woods Hole, Massachusetts 02543.
Telephone: (508) 548-3705. ext. 428. FAX: 508-540-6902. E-mail: pamcl@hoh.mbl.edu.
POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN. Marine Biological Laboratory,
Woods Hole, MA 02543.
Copyright © 1992, by the Marine Biological Laboratory
Second-class postage paid at Woods Hole, MA, and additional mailing offices.
ISSN 0006-3 185
INSTRUCTIONS TO AUTHORS
The Biological Bulletin accepts outstanding original research
reports of general interest to biologists throughout the world.
Papers are usually of intermediate length (10-40 manuscript
pages). A limited number of solicited review papers may be ac-
cepted after formal review. A paper will usually appear within
four months after its acceptance.
Very short, especially topical papers (less than 9 manuscript
pages including tables, figures, and bibliography) will be pub-
lished in a separate section entitled "Research Notes." A Re-
search Note in The Biological Bulletin follows the format of
similar notes in Nature. It should open with a summary para-
graph of 150 to 200 words comprising the introduction and the
conclusions. The rest of the text should continue on without
subheadings, and there should be no more than 30 references.
References should be referred to in the text by number, and
listed in the Literature Cited section in the order that they appear
in the text. Unlike references in Nature, references in the Re-
search Notes section should conform in punctuation and ar-
rangement to the style of recent issues of The Biological Bulletin.
Materials and Methods should be incorporated into appropriate
figure legends. See the article by Lohmann et al. (October 1990,
Vol. 179: 214-2 1 8) for sample style. A Research Note will usually
appear within two months after its acceptance.
The Editorial Board requests that regular manuscripts con-
form to the requirements set below; those manuscripts that do
not conform will be returned to authors for correction before
review.
1. Manuscripts. Manuscripts, including figures, should
be submitted in triplicate. (Xerox copies of photographs are not
acceptable for review purposes.) The original manuscript must
be typed in no smaller than 12 pitch, using double spacing (in-
cluding figure i 'ends, footnotes, bibliography, etc.) on one side
of 16- or 20-lb. I oaper, 8'/2 by 11 inches. Please, no right
justification. Manuv. i should be proofread carefully and er-
rors corrected legib.s . , ,-k ink. Pages should be numbered
consecutively. Margins on al1 sides should be at least 1 inch (2.5
cm). Manuscripts should conform to the Council of Biology Ed-
itors Style Manual, 5th Edition (Council of Biology Editors.
1983) and to American spelling. Unusual abbreviations should
be kept to a minimum and should be spelled out on first referenc
as well as defined in a footnote on the title page. Manuscript,
should be divided into the following components: Title page
Abstract (of no more than 200 words). Introduction, Materials
and Methods, Results, Discussion. Acknowledgments. Literature
Cited, Tables, and Figure Legends. In addition, authors should
supply a list of words and phrases under which the article should
be indexed.
2. Title page. The title page consists of: a condensed title
or running head of no more than 35 letters and spaces, the
manuscript title, authors' names and appropriate addresses, and
footnotes listing present addresses, acknowledgments or contri-
bution numbers, and explanation of unusual abbreviations.
3. Figures. The dimensions of the printed page, 7 by
inches, should be kept in mind in preparing figures for publ
cation. We recommend that figures be about 1 '/: times the line;1
dimensions of the final printing desired, and that the ratio c
the largest to the smallest letter or number and of the thickest
to the thinnest line not exceed 1:1.5. Explanatory matter gen-
erally should be included in legends, although axes should always
be identified on the illustration itself. Figures should be prepared
for reproduction as either line cuts or halftones. Figures to be
reproduced as line cuts should be unmounted glossy photo-
graphic reproductions or drawn in black ink on white paper,
good-quality tracing cloth or plastic, or blue-lined coordinate
paper. Those to be reproduced as halftones should be mounted
on board, with both designating numbers or letters and scale
bars affixed directly to the figures. All figures should be numbered
in consecutive order, with no distinction between text and plate
figures. The author's name and an arrow indicating orientation
should appear on the reverse side of all figures.
4. Tables, footnotes, figure legends, etc. Authors should
follow the style in a recent issue of The Biological Bulletin in
preparing table headings, figure legends, and the like. Because
of the high cost of setting tabular material in type, authors are
asked to limit such material as much as possible. Tables, with
their headings and footnotes, should be typed on separate sheets,
numbered with consecutive Roman numerals, and placed after
he Literature Cited. Figure legends should contain enough in-
irmation to make the figure intelligible separate from the text,
egends should be typed double spaced, with consecutive Arabic
jmbers. on a separate sheet at the end of the paper. Footnotes
lould be limited to authors' current addresses, acknowledg-
ments or contribution numbers, and explanation of unusual
bbreviations. All such footnotes should appear on the title page.
rootnotes are not normally permitted in the body of the text.
5. Literature cited. In the text, literature should be cited
by the Harvard system, with papers by more than two authors
riled as Jones et ai, 1980. Personal communications and ma-
erial in preparation or in press should be cited in the text only,
/ith author's initials and institutions, unless the material has
:en formally accepted and a volume number can be supplied,
he list of references following the text should be headed Lit-
ature Cited, and must be typed double spaced on separate
ages, conforming in punctuation and arrangement to the style
f recent issues of The Biological Bulletin. Citations should in-
clude complete titles and inclusive pagination. Journal abbre-
/iations should normally follow those of the U. S. A. Standards
Institute (USASI). as adopted by BIOLOGICAL ABSTRACTS and
CHEMICAL ABSTRACTS, with the minor differences set out below .
The most generally useful list of biological journal titles is that
published each year by BIOLOGICAL ABSTRACTS (BIOSIS List of
Serials: the most recent issue). Foreign authors, and others who
are accustomed to using THE WORLD LIST OF SCIENTIFIC PE-
RJODICALS, may find a booklet published by the Biological
Council of the U.K. (obtainable from the Institute of Biology,
41 Queen's Gate, London, S.W.7, England, U.K.) useful, since
it sets out the WORLD LIST abbreviations for most biological
journals with notes of the USASI abbreviations where these differ.
CHEMICAL ABSTRACTS publishes quarterly supplements of ad-
itional abbreviations. The following points of reference style
or THE BIOLOGICAL BULLETIN differ from USASI (or modified
WORLD LIST) usage:
A. Journal abbreviations, and book titles, all underlined
(for italics)
B. All components of abbreviations with initial capitals
(not as European usage in WORLD LIST e.g.. J. Cell. Comp.
Physiol. NOT / cell. comp. Physiol.)
C. All abbreviated components must be followed by a pe-
riod, whole word components must not (i.e., J. Cancer Res.)
D. Space between all components (e.g.. J. Cell. Comp.
Physiol.. not J.Cell.Comp.Physiol.)
E. Unusual words in journal titles should be spelled out
in full, rather than employing new abbreviations invented by
the author. For example, use Rit I 'fsindajjelags Islendinga with-
out abbreviation.
F. All single word journal titles in full (e.g., Veliger. Ecol-
ogy. Brain).
G. The order of abbreviated components should be the
same as the word order of the complete title (i.e., Proc. and
Trans, placed where they appear, not transposed as in some
BIOLOGICAL ABSTRACTS listings).
H. A few well-known international journals in their pre-
ferred forms rather than WORLD LIST or USASI usage (e.g.,
Nature, Science, Evolution NOT Nature. Land.. Science, N.Y.;
Evolution, Lancaster, Pa.)
6. Reprints, page proofs, and charges. Authors receive their
first 100 reprints (without covers) free of charge. Additional re-
prints may be ordered at time of publication and normally will
be delivered about two to three months after the issue date.
Authors (or delegates for foreign authors) will receive page proofs
of articles shortly before publication. They will be charged the
current cost of printers' time for corrections to these (other than
corrections of printers' or editors' errors). Other than these
charges for authors' alterations. The Biological Bulletin does not
have page charges.
Hi
ERRATUM
The Biological Bulletin, Volume 180, Number 2, page 314
The following correction should be made in the article by Anthony Pires and Michael G. Hadneld titled,
"Oxidative breakdown products of catecholamines and hydrogen peroxide induce partial metamorphosis in
the nudibranch Phestilla sibogae Bergh (Gastropoda: Opisthobranchia)" (Biol. Bull. 180: 310-317).
On page 314, the second sentence of the second paragraph in the left hand column, which reads "Con-
centration threshold for velar loss after 7-h exposure to fresh DA. . ." should read, "Concentration threshold
for velar loss after 7-h exposure to aged DA. . ." The word "aged" replaces the word "fresh."
Reference: Biol. Bull. 182: 1-7. (February, 1992)
Fast-Strike Feeding Behavior in a Pteropod Mollusk,
Clione limacina Phipps
COLIN O. HERMANS' AND RICHARD A. SATTERLIE
Department of Zoology, Arizona State University. Tempe, Arizona 85287
and Friday Harbor Laboratories. Friday Harbor. Washington 98250
Abstract. High speed cinematography and video re-
cordings were used to evaluate the fast-strike feeding re-
sponse by which Clione limacina captures its prey, Li-
macina fielicina. The acquisition phase of feeding involves
rapid mouth opening and extrusion of three pairs of buccal
cones. Mouth opening occurs in 10 to 20 ms, while hy-
drostatic inflation of the buccal cones takes 50 to 70 ms.
Buccal cones are immediately retracted if prey are not
contacted. Buccal cones surround the prey and release a
viscous material that may be used as an adhesive attach-
ment to the prey shell. Surface ultrastructure of the buccal
cones reveals that they are studded with clusters of capit-
ulate papillae, which appear to be the source of the viscous
secretory material.
Introduction
The pteropod mollusk Clione limacina feeds exclusively
on shelled pteropods (Lalli and Gilmer, 1989). Due to
the extremely limited dietary breadth of Clione, as well
as the active swimming characteristics of both predator
and prey (Limacina helicina in Friday Harbor, Washing-
ton), it is not surprising to find rapidly activated and highly
specialized feeding structures in Clione. For prey acqui-
sition, Clione rapidly extrudes three pairs of oral append-
ages, called buccal cones, which surround and adhere to
the prey (see Lalli, 1970). Each buccal cone is cone shaped
when retracted, but becomes more cylindrical when ex-
truded. Extrusion of buccal cones is primarily due to hy-
draulic inflation (Lalli and Gilmer, 1989). The acquisition
phase of feeding is followed by a manipulative phase, dur-
ing which the prey is turned so that the shell opening is
Received 5 June 1991; accepted 25 November 1991.
' Present address: Department of Biology, Sonoma State University,
Rohnert Park. California 94928.
over the mouth of Clione. Manipulation is performed by
the buccal cones and is followed by the consumptive
phase, during which the prey is extricated from its shell.
Extrication involves the use of two specialized hook sacs
that form part of the buccal apparatus (Lalli, 1970). Each
hook sac contains tufts of recurved chitinous hooks, which
are protracted into the shell opening to grasp and pull the
prey from its shell. Soft tissues of the prey are dislodged
by alternate protractions and retractions of the hook sacs.
Swallowing is aided by protraction and retraction move-
ments of the radula, which is also part of the buccal ap-
paratus. The soft tissues of the prey are swallowed whole
(Wagner, 1885; Litvinova and Orlovsky, 1985; for other
references see Lalli and Gilmer, 1989).
Two distinct forms of feeding behavior are observed.
In the first, referred to here as the fast-strike response,
Clione enters the acquisition phase of feeding from an
unexcited, slow swimming activity state. During acqui-
sition, swimming changes from slow to fast, and continues
fast throughout the consummatory phase. During fast
swimming, bending of the tail leads to frequent turning
and looping movements of the entire body. If a fast-strike
fails, and prey is not acquired, the buccal cones are im-
mediately withdrawn, and fast swimming is terminated.
The fast-strike response, which is initiated by prey contact,
thus represents a sudden change to feeding behavior; if
unsuccessful, the response is terminated by an equally
sudden return to pre-strike swimming activity.
The second type of feeding behavior is initiated without
direct physical contact with the prey. This activity involves
fast swimming with loops and turns, as well as buccal
cone extrusion and is referred to as "hunting behavior"
(Litvinova and Orlovsky, 1985). Hunting behavior can
be induced by placing an animal in seawater containing
prey homogenates, by placing non-feeding Clione indi-
viduals close to feeding individuals, or by injecting sero-
C. O. HERMANS AND R. A. SATTERLIE
tonin into the hemocoel (Litvinova and Orlovsky, 1985;
Kabotyanski and Sakharov, 1988). Hunting behavior is
similar to fast-strike feeding behavior in that the mouth
is held open with the buccal cones protruding, and swim-
ming is fast with frequent changes in direction. The be-
haviors differ in two important ways. First, hunting be-
havior does not require direct contact with an intact prey.
Second, buccal cone extension and fast swimming are
maintained in hunting behavior, whereas both are ter-
minated immediately in the fast-strike if a prey item is
not acquired. Note that the prey acquisition responses of
Clione form a continuum, with fast-strike feeding at one
extreme and indefinite hunting behavior at the other.
In this paper, we describe behavioral and morphological
aspects of the acquisition phase of fast-strike feeding; a
cine analysis of mouth and buccal cone movements and
a description of the surface morphology of the buccal
cones are included. This work provides the background
for an ongoing electrophysiological investigation into the
acquisition phase of feeding behavior and the role of pu-
tative modulators on the motivational states underlying
feeding behavior. It also extends the cinematic analysis
of Clione feeding behavior by Litvinova and Orlovsky
(1985).
Materials and Methods
Both Limacina and Clione were collected from the
breakwater at Friday Harbor Laboratories, Friday Harbor,
Washington, and held in one-gallon glass jars in a seawater
table. Individual animals were filmed in a small glass
chamber filled with seawater at room temperature (16-
18°C). Fast-strike sequences were filmed, within five days
of animal collection, at 100 frames/s with a Redlake Lo-
cam high speed 16 mm camera containing Kodak Plus-
X negative film. Additional feeding sequences were
"filmed" with a Sony CCD video camera HVM-200,
equipped with a Nikon Micro-Nikkor lens, at the equiv-
alent of 60 frames/s and were recorded on a Canon VR-
30 4-head portable video recorder. Feeding sequences were
obtained by touching the prey, Limacina helicina, to
swimming individuals of Clione. Limacina were attached
with "Super Glue" to a human head hair or held in fine
forceps.
A hair was attached to the Limacina shell as follows.
A Limacina was placed in a shallow container on the
stage of a dissecting microscope and the water level in the
container was lowered until the shell, which is very hy-
drophobic, broke through the surface film of the water.
The Limacina was then turned to achieve the desired ori-
entation. The root of a human hair was quickly dipped
in a small droplet of "super glue" and applied to the sur-
face of the shell.
Fast-strike responses were recorded from five different
individuals. One complete response (from initiation
through acquisition) was recorded from each of these an-
imals, but unsuccessful strikes were often recorded before
the complete event. Unsuccessful strikes were also re-
corded from three other individuals that never produced
a complete response. All animals were between 1 .4 and
2.2 cm in body length.
Film sequences were analyzed frame-by-frame by
making photographic prints of the sequences, and by pro-
jecting individual frames onto tracing paper. Tracings
were made of body, wing, head, and buccal cone positions.
In one case, the images from sequential frames were dig-
itized from tracings with a Jandel Scientific digitizing pad
and processed with a computer-assisted three-dimensional
reconstruction software program, (PC3D™, Jandel Sci-
entific, Corte Madera, California). Photographic prints
were made by projecting 16-mm frames directly onto
photographic paper with a standard photographic enlarger.
Video sequences, advanced frame-by-frame, were traced
directly from a television screen during viewing.
For scanning electron microscopical investigation,
specimens that were not adhering to prey were anesthe-
tized by immersion in a 1 : 1 solution of 0.33 A/ magnesium
chloride and seawater. A Clione adhering to its prey was
prepared as follows. First, a Limacina, glued to a hair,
was dangled in an aquarium so as to contact swimming
individuals of Clione. When one of the pteropods struck
at and gripped the prey, it was immediately pulled out of
the aquarium and dropped directly into the primary fix-
ative solution. The specimen continued to grip its prey
as they were both being fixed, and remained attached until
CO2 turbulence, during critical point drying, accidentally
separated them, exposing the adherent surfaces. Fixation
was completed by immersion in isotonic, cacodylate-buf-
fered 2% glutaraldehyde, pH 7.3, at room temperature
for 2 h, and postfixation was in cacodylate-buffered 1%
osmium tetroxide for 1 h at room temperature. The spec-
imens were dehydrated in ethanol, critical point dried
from carbon dioxide, and sputter coated with gold and
palladium before examination with an AMRay 1000 (Figs.
1, 4A) or a JEOL JSM-35 (Fig. 4B) scanning electron
microscope.
Results
Acquisition behavior
Fast-strike feeding behavior was initiated by bringing
a tethered Limacina into contact with the oral region of
a freely swimming Clione. In our experience, the success
rate of inducing fast-strikes was extremely low. With some
animals, a day or more would pass without a strike being
elicited; Clione apparently feeds irregularly. The degree
of satiation in individual animals could not, therefore, be
determined. The success rate was equally low, however,
in animals that had been held in a jar for more than a
FAST-STRIKE FEEDING
week. With other animals, strikes could he obtained with
some dependability. On one occasion, a response was ob-
tained although the prey was not in contact with the oral
region of Clione. In this case, the Limacina began rapid
swimming movements when brought near the oral region
of Clione, triggering an immediate fast-strike response.
In all observed fast-strike responses, the initial response
of the acquisition phase was rapid mouth opening. When
closed, the mouth forms a dorsoventral slit on the anterior
margin of the head (Fig. 1 A). Lip retraction pulls the lips
laterally, causing mouth gaping and protrusion of the
buccal cones (Fig. IB). The degree of mouth opening can
be judged from the position of the anterior tentacles, as
recorded on film and video prior to and during fast-strike
responses (Figs. 2, 3). The mouth of Clione is flanked by
a pair of anterior tentacles that project from the antero-
lateral margins of the head (Fig. 1 A). When Clione is hov-
ering or slowly swimming forward (upward), the anterior
tentacles are normally inflated and project forward (Figs.
1A, 2 A). During mouth opening, lip retraction, and pro-
trusion of the buccal cones, the anterior tentacles rotate
laterally 90°, so that their projection is perpendicular to
the longitudinal axis of the animal (Figs. 2D, 3). Mouth
opening occurs in the first 20 ms of the fast-strike and is
accompanied by full exposure and partial protraction of
the buccal cones (Fig. 3). This can be demonstrated by
pulling open the mouth of an anesthetized animal, which
exposes the buccal cones and causes them to bulge slightly
out of the mouth (similar to that seen in Fig. IB). Three
buccal cones lie on either side of the buccal mass (a mus-
cular organ containing the radula and a pair or hook sacs),
in a line parallel to the lips. The retracted cones are not
inverted, but rather are collapsed and retracted into small
cavities, or cheek pouches, adjacent to the buccal mass.
Buccal cones protract when they are inflated with he-
molymph (Lalli and Gilmer, 1989). This is supported by
our physiological experiments in which induced activity
in buccal cone protraction motor neurons causes mouth
opening, contraction of head musculature, but only partial
extension of buccal cones (Norekian and Satterlie, in
prep.). In these preparations, full expansion of the buccal
cones is impossible because the head hemocoel is com-
promised to allow electrophysiological recordings. In in-
tact animals, expanded cones can extend approximately
one-half body length from the mouth. Expansion is ac-
companied by a decrease in the diameter of the head and
the appearance of a distinct circular constriction in the
neck region (Figs. 2D, 3). In two recorded sequences in
which the head and neck outlines were clearly shown, the
reduction in head diameter averaged 22.7% while the re-
duction in neck diameter averaged 20.2%. Full expansion
of the buccal cones, including the initial mouth opening,
takes from 50 to 70 ms (Fig. 3). If the prey is not contacted
during buccal cone expansion, the cones are immediately
retracted, the mouth is closed, and the animal returns to
slow swimming. Retraction of buccal cones is not a passive
deflation, because the cones can be fully retracted in 70
to 90 ms (based on three unsuccessful strikes). On two
occasions, strikes were aborted when the buccal cones were
inflated to only 10 to 20% of the body length. In these
cases, the cones were immediately retracted as in unsuc-
cessful strikes.
Inflation of the buccal cones occurs from the base out-
ward; the tips of the cones do not inflate until late in cone
expansion. The uninflated tips are more opaque than the
inflated parts of the buccal cones (Fig. 2D). As the cones
Figure 1. Scanning electron micrographs showing ventral views of heads of Clione in normal swimming
posture ( 1 A) and with mouth (m) open and five of six buccal cones (be) partially protruded ( 1 B). Note the
pair of anterior tentacles (t) that bear ciliary tufts (c). The head (h) is covered with a coat of motile cilia,
w — wings, Ic — tufts of large neck cilia.
C. O. HERMANS AND R. A. SATTERLIE
Figure 2. Representative frames from cinematographic series taken at 100 frames/s showing a tethered
Limacina being offered to a Clione. cw — wings of Clione. wl — wings of Limacina. be — buccal cones, t —
anterior tentacles of Clione. (A) Predator and prey 200 ms (20 frames) before first sign of response to contact.
(B) First sign of response to contact. Note the slight bulge on head of Clione (arrow). (C) Next frame (10
ms) after (B), showing buccal cones exploding from cheek pouches and forming grasping tentacles. (D) 4
frames (40 ms) after (C), showing buccal cones near full extension and beginning to grip the Limacina. Note
the decreased diameter of the head, and the prominent neck constriction.
are extruded, they project outward at approximately 45°
with a slight concave curvature with respect to the mouth.
As the cones reach full expansion, they bend around the
prey and adhere to its shell (Fig. 3).
Limacina shells pulled from the grasp of Clione buccal
cones were coated with a viscous residue in the regions
contacted by buccal cones. Clear viscous material pro-
duced by the buccal cones could be gripped with fine for-
ceps and lifted in fine strands from the surface of the sea-
water containing the Clione. During hunting behavior,
the protracted buccal cones frequently adhered to the wall
of the container following contact with it. Removal of an
adhering animal revealed residue on the glass, apparently
adhesive.
Surface ultraslnicture of the buccal cones
The surface of each buccal cone is studded with clusters
or rosettes of capitulate papillae (Fig. 4). The number of
papillae in each cluster varies from two or three to about
a dozen. The clusters near the bases of the buccal cones
contain the fewest papillae per rosette, those toward the
tips contain more. Each papilla is about 1 5 nm high and
somewhat less than 10 ^m in diameter. The tip of each
papilla is slightly inflated, forming a lumpy capitulum
about 10 //m in diameter. Each rosette has a common
stalk, about 20 /urn in diameter and 20 yum in height. Long
cilia protrude from the sides of the papillae and project
from the surface of the buccal cone between the papillae.
Tight clusters of cilia protrude from the centers of some
of the papillary rosettes. Isolated clusters of cilia, were
also observed, but they were not common (Fig. 4A).
When the buccal cones are retracted, the epidermis be-
tween the clusters of papillae is deeply folded, and the
capitula and cilia form a tightly packed feltwork or welter
on the surface of each cone. When the buccal cones are ex-
tended, the rosettes of papillae stand up above a smooth,
simple squamous epithelium that stretches tightly over the
tentacular surface between the rosettes of papillae (Fig. 4A).
The surface of the shell of Limacina, to which the buccal
cones adhere, is very smooth, transparent, and very hy-
drophobic; it appears smooth when viewed with a scan-
ning electron microscope. The shells of dead Limacina
lose their hydrophobic properties rapidly. Where the shell
of a Limacina is contacted by the buccal cones of a fast-
striking Clione are fine threads, observable by SEM, that
correspond to those that appear on the surfaces of the
buccal cones where they contact the Limacina shell (Fig.
4B). These threads appear to originate from the tips of
the capitulate papillae, but this possibility is difficult to
establish with certainty.
Discussion
The fast-strike response of Clione consists of a rapid
opening of the mouth and a hydraulic inflation of the six
buccal cones, the entire response occurring in 50 to 70
ms. In aborted or unsuccessful strikes, withdrawal of buc-
FAST-STRIKE FEEDING
Figure 3. Tracings of Clione and Limacina from cine series with
time intervals of 10ms between frames and covering the 100 ms interval
from one frame (10 ms) prior to the first sign of response to the prey
through the initial grasping of the shell. The sequence has been plotted
twice with a 7° shift in the y-axis. When viewed with a stereoscopic
viewer or with crossed eyes, the sequence will appear in 3-dimensions
with time represented in the z-plane. Buccal cone labels: (Id) — left dorsal,
(1m) — left median, (Iv) — left ventral, (rd) — right dorsal, (rm) — right me-
dial. Right (rat) and left (lat) anterior tentacles are also labelled.
cal cones is nearly as rapid. This would suggest that both
expansion and withdrawal are active responses. Fast-strike
prey acquisition is thus distinct from the hunting behavior
described by Litvinova and Orlovsky (1985) in which
Clione rapidly swim with the buccal cones held in an ex-
panded state. The initial phase of hunting behavior pre-
sumably involves similar mouth opening and buccal cone
inflation.
The low success rate in triggering a fast-strike under
laboratory conditions suggests that the fast-strike response
has a high threshold for activation. Lowering of this
threshold could result in behavior that is more disposed
toward feeding, such as hunting behavior. In this case,
the difference between responses to prey during normal
swimming and those during hunting behavior might be
one of motivational state. This difference can best be il-
lustrated by comparing buccal cone responses during
hunting and during an unsuccessful fast-strike response.
In the former, the buccal cones are held in an expanded
position despite the lack of mechanical contact with the
prey. In the latter case, lack of prey contact results in a
rapid withdrawal of the buccal cones and a return to nor-
mal swimming behavior. In hunting behavior, therefore,
buccal cone withdrawal must be suppressed, even in the
absence of direct mechanical contact with prey.
The nature of the trigger underlying the change in be-
havioral state, from hunting to fast-strike, is not known.
It may, however, involve serotonergic input to the feeding
system, because bath application or hemocoel injection
of serotonin can trigger behavioral responses similar to
those of hunting behavior; i.e., the responses can be
evoked although the animal has not been exposed to prey
or prey extracts (Kabotyanski and Sakharov, 1988). The
external stimulus for a switch to hunting behavior pre-
sumably involves chemosensory input because Limacina
extracts, or proximity to feeding Clione, can initiate hunt-
ing behavior (Litvinova and Orlovsky, 1985).
Inflation of the buccal cones is remarkable for its
great speed. Expansion is associated with a decrease in
head and neck diameter, suggesting that increased
hemocoelic pressure is associated with buccal cone in-
flation. Arshavsky et al. (1990) have shown that heart
rate in Clione increases during hunting behavior, further
supporting the idea that feeding responses are associated
with increases in hemocoelic pressure. Pressure changes
can be localized in the head as a muscular diaphragm
separates head and body hemocoels. The diaphragm
surrounds the anterior aorta and may act as a physio-
logical valve further regulating blood flow to the hem-
ocoel in the head (Lalli, 1967).
With buccal cones protruded, the Clione appear much
like small squid. This led Wagner (1885) and Pelseneer
(1885) to consider the possibility of homology between
Clione buccal cones and squid tentacles. However, em-
bryonic origins and innervation patterns demonstrate that
they are not homologous (see Lalli and Gilmer, 1989. for
a discussion of pteropod systematics and affinities). The
mechanisms by which the two types of tentacles move to
grasp their prey are quite distinct. Kier demonstrated that
cephalopod tentacles are muscular hydrostats (Kier, 1982,
1987, 1988; Smith and Kier, 1989). Muscular hydrostats
are readily distinguishable from hydrostatic skeletons that
use a hydraulic mechanism in that their volumes are made
up almost entirely of muscular tissue. Therefore, although
they can undergo extensive changes in shape, muscular
hydrostats do not substantially change volume. Hydraulic
hydrostatic skeletons, in contrast, are fluid-filled cavities
surrounded by muscular or fibrous tissues that resist the
hydrostatic pressure within (Smith and Kier, 1989).
No clear differences are found when the speed of ten-
tacle protractions in muscular hydrostats is compared with
that of the hydraulic system of Clione, because the range
of protraction speeds found in muscular hydrostat systems
is very wide. For example, each of the 19 pairs of digital
tentacles of Nautilus consists of an extensible, muscular.
C. O. HERMANS AND R. A. SATTERLIE
Figure 4. (A) Enlarged view of a part of one of the partially protruded buccal cones shown in Figure
IB. The head of each papilla (p) in the rosettes is studded with bumps. Motile cilia (c) project from the shaft
of each papilla, whereas tufts of cilia (sc) project from the centers of rosettes, or less commonly are isolated
from the rosettes. (B) Similar view of a region on a buccal cone of a different specimen, which was allowed
to adhere to a Limacina shell and was fixed while grasping the prey. Dense mats of thread-like structures
(t) can be seen on the adherent surfaces of buccal cones. Some appear to originate from the tips of capitulate
papillae (arrows).
adhesive cirrus enclosed in a protective sheath. Protrusion
of the cirrus from the tip of its sheath, which is necessary
for it to grasp prey, requires 5-10 s or longer (Kier, 1987).
At the opposite extreme, the tentacles of squid elongate
fully in 15 to 30 ms (Keir, 1982, 1985).
The buccal cones of Clione can be protruded in less
than 100 ms, and this performance is best appreciated
when compared to other fast invertebrate prey capture
behaviors that have been subjected to cine analysis. For
example, prey seizure in the opisthobranch mollusk Na-
vanax occurs in 380 ms; this is a muscular phenomenon
involving a pharyngeal lunge followed by lip closure
around the prey (Susswein el a/.. 1984; Susswein and
Achituv, 1987). Prey acquisition behaviors involving the
movement of body parts that are supported by hard skel-
etal elements can be much faster; e.g., the strike of the
second thoracic appendages of stomatopod crustaceans
occurs in 4-8 ms (Burrows, 1969).
Other gastropods can strike rapidly. In particular, the
proboscides of toxoglossans, which contain poisonous.
dart-like radular teeth, are potentially as rapid as the
Clione buccal cone system. Predatory strikes have been
described and photographed in a turrid, Ophiodermella
inermis (Shimek and Kohn, 1981), and in Conns (Ny-
bakken, 1967). The proboscis of Conns is protruded as a
hydrostatic skeleton (Greene and Kohn, 1989), but the
speed with which these strikes occur has not been analyzed
by high speed cine or video.
Whereas some gymnosomatous pteropods do appre-
hend their prey with suction cups, somewhat like coleo-
idean cephalopods (Lalli and Gilmer, 1989; Kier and
Smith, 1990), the adhesiveness of the buccal cones of
Clione resembles that of the digital tentacles of Nautilus.
The digital tentacles grip prey by means of ridges on the
cirri that protrude from the tips of the sheaths that form
the bases of the tentacles (Fukuda, 1987; Kier, 1987). In
both Nautilus and Clione, the adhesive structures are en-
sheathed when not in use. In both cases, a question re-
mains: to what degree are the prey simply gripped, and
to what extent do the tentacles adhere? Fukuda (1987)
FAST-STRIKE FEEDING
suggested that the ensheathing of the cirri in Nautilus
might serve to save mucus.
Apprehension of the prey by Clione may be partly by
chemical adhesion and partly by the physical gripping of
the Limacina shell by the enclosing buccal tentacles. The
capitula of the papillae on the buccal tentacles might be
thrust through the boundary layer of water, covering the
hydrophobic surface of the Limacina shell and providing
the means of attachment to, or gripping of the shell, just
like the beaded gloves used by soccer goalies and football
wide receivers aid in gripping the wetted, hydrophobic
surfaces of footballs. In both cases, the bumps aid in
adhesion; they penetrate the boundary layer of water,
eliminating this weak boundary layer by driving it into
the spaces between the bumps (Waite, 1987).
Because the buccal tentacles appear to be chemically
adhesive and yet can detach to manipulate the shell of
the prey so that the opening is aligned with the Clione
mouth, the possibility that both adhesive and releasing
chemicals are secreted must be considered (Hermans,
1983). Examination of the ultrastructure of the buccal
cones and their secretions, as well as analyses of the control
of feeding behavior, will help answer this and other ques-
tions about prey acquisition in Clione.
Acknowledgments
We thank Dr. Tigran Norekian for translating Litvinova
and Orlovsky (1985), Ms. Michelle Lagro for preparing
specimens of Clione for electron microscopy. Prof.
A. O. D. Willows of Friday Harbor Laboratories for space
and equipment. Prof. R. Strathmann for use of his cine
camera and the macro lens and video equipment. Dr.
Tom Schroeder for instruction in SEM, also Mr. W. Sharp
for instruction in EM and for the use of the Biological
Electron Microscope Facility at Arizona State University,
Mr. Chaz. Kazelik for help with the PC3D stereoscopic
imaging program, Drs. Claudia Mills and Norm McLean
for collecting and shipping specimens, and several other
friends, colleagues, staff, and family at the Friday Harbor
Labs for help in collecting specimens and for many other
kindnesses. Thanks also to Sarah Cohen for suggesting
the use of "Super Glue." Our perspective on the potential
similarities between the strikes of toxoglossans and Clione
has benefitted from discussions with Drs. Ron Shimek,
Matt James, and Ed Smith.
Literature Cited
Arshavsky, Y. I., T. G. Deliagina, I. M. Gelfand, G. N. Orlovsky, Y. V.
Panchin, G. A. Pavlova, and L. B. Popova. 1990. Neural control
of heart beat in the pteropod mollusc Clione limacina: coordination
of circulatory and locomotor systems. J. Exp. Biol. 148: 461-475.
Burrows, M. 1969. The mechanics and neural control of the prey cap-
ture strike in the mantid shrimps Stiuilla and Hemisquilla. Z I'ergl.
Physiol.62: 361-381.
Fukuda, Y. 1987. Histology of the long digital tentacles. Pp. 249-256
in Nautilus: The Biology and Paleobiology of a Living Fossil. W. B.
Saunders and N. H. Landman. eds. Plenum, New York.
Greene, J. L., and A. J. Kohn. 1989. Functional morphology of the
Conns proboscis (Mollusca: Gastropoda). J Zoo/., Loud. 219: 487-
493.
Hermans, C. O. 1983. The duo-gland adhesive system. Oceanogr. Mar.
Biol. Ann. Rev. 21:283-339.
Kabotyanski, E. A., and D. A. Sakharov. 1988. Monoamine-dependent
behavioural states in the pteropod mollusc Clione limacina. Symposia
Biologica Hungarica 36: 463-477.
Kier, W. M. 1982. Functional morphology of the musculature of squid
(Loliginidae) arms and tentacles. J. Morphol. 172: 179-192.
Kier, W. M. 1985. The musculature of squid arms and tentacles: ul-
trastructural evidence for functional differences. / Morphol. 185:
223-239.
Kier, W. M. 1987. The functional morphology of the tentacle mus-
culature of Nautilus pompilius. Pp. 257-269 in Nautilus: The Biology
and Paleobiology of a Living Fossil. W. B. Saunders and N . H . Land-
man. eds. Plenum. New York.
Kier, W. M. 1988. The arrangement and function of molluscan muscle.
Pp. 21 1-252 in The Mollusca. Form and Function. Vol. 1 1. E. R.
Trueman and M. R. Clarke, eds. Academic Press, New York.
Kier, W. M., and A. M. Smith. 1990. The morphology and mechanics
of octopus suckers. Biol. Bull. 178: 126-136.
Lalli, C. M. 1967. Studies on the structure and biology of two gym-
nosomatous pteropods, Clione kincaidi Agersborg and Crucibran-
chaea macrochira (Meisenheimer). Ph.D. Dissertation, University of
Washington. 175 pp.
Lalli, C. M. 1970. Structure and function of the buccal apparatus of
Clione limacina (Phipps) with a review of feeding in gymnosomatous
pteropods. J. Exp. Mar. Biol. Ecol. 4: 101-1 18.
Lalli, C. M., and R. W. Gilmer. 1989. Pelagic Snails: The Biology of
Holoplanktonic Gastropod Mollusks. Stanford University Press,
Stanford. 259 pp.
Litvinova, N. M., and G. N. Orlovsky. 1985. Feeding behavior of the
pteropod mollusc Clione limacina. Byull. MOIP. Old. hiol. 90: 73-
77.
Nybakken, J. 1967. Preliminary observations on the feeding behavior
of Conus purpurascens Broderip, 1833. I'e/iger 10: 55-57.
Pelseneer, P. 1885. The cephalic appendages of the gymnosomatous
pteropoda, and especially ofC/ione. Q. J. Microsc. Sci. 25: 491-509.
Shimek, R. L., and A. J. Kohn. 1981. Functional morphology and
evolution of the toxoglossan radula. Malacologia 20: 423-438.
Smith, K. K., and W. M. Kier. 1989. Trunks, tongues, and tentacles:
moving with skeletons of muscle. Am. Sci. 77(1): 28-35.
Susswein, A. J., and Y. Achituv. 1987. Pharyngeal movements during
feeding sequences ofNavanax inermis (Gastropoda: Opisthobranchia)
in successive stages of dissection. / Exp. Biol. 128: 323-333.
Susswein, A. J., Y. Achituv, M. S. Cappell, D. C. Spray, and M. V. L.
Bennett. 1984. Pharyngeal movements during feeding sequences
in Navanax inermis: a cinematographic analysis. J. Comp. Physiol.
155:209-218.
Wagner, N. 1885. Die H'irbcllosen des Weissen Meeres. Wilhelm En-
gelmann, Leipzig. Pp. 1-168.
Waite, J. H. 1987. Nature's underwater adhesive specialist. Inl. J
Adhesion and Adhesives 1: 9-14.
Reference: Biol. Bull. 182: 8-14. (February, 1992)
Effects of Photoperiod and Temperature on
Egg-Laying Behavior in a Marine Mollusk,
Aplysia californica
NANCY L. WAYNE AND GENE D. BLOCK
Department of Biology, University of Virginia, C/iarlottesville, Virginia 22901
Abstract. The primary purpose of these studies was to
determine whether photoperiodic signals could influence
seasonal egg-laying behavior in the marine mollusk, Aply-
sia californica. Egg-laying behavior was monitored from
groups of animals that were collected at four times of year
and maintained in different temperature and photoperi-
odic conditions in the laboratory. Animals that were ob-
tained in autumn and kept in warm water laid eggs more
frequently than those in cold water, regardless of photo-
period. Furthermore, animals maintained on short days
and warm water laid eggs more frequently than those on
long days and warm water. Animals in cold water showed
little to no egg laying, and a photoperiodic response was
not evident. Animals that were collected in either winter
or spring and maintained in warm water showed little or
no spontaneous egg laying throughout the study, regardless
of photoperiod. As with the autumn animals, Aplysia in-
dividuals obtained in summer and kept on short days and
warm water laid eggs more frequently than those kept on
long days and warm water. These results provide the first
evidence that the reproductive system of A californica is
responsive to photoperiod. Overall, the data suggest that
warm water is permissive for egg laying, and that short
days can further stimulate this behavior. However, there
is a strong inhibition of spontaneous egg laying during
the winter and spring, which neither warm water nor short
photoperiod can overcome. The role of the eyes in me-
diating the photoperiodic response was also investigated.
A control group of intact animals kept on short days laid
eggs more frequently than those on long days, but this
photoperiodic response was not evident in eyeless
Received 12 August 1 99 1 ; accepted 31 October 1991.
animals. These results suggest that the eyes play a role
in mediating the effects of photoperiod on egg laying
behavior.
Introduction
Like many animals living in the temperate zone, the
marine mollusk Aplysia californica breeds seasonally.
Both field and laboratory observations indicate that this
species is reproductively competent during the summer
and autumn, and reproductively quiescent during the
winter and spring (Strumwasser el al. 1969; Audesirk,
1979; Berry, 1982). The onset of the breeding season is
indicated by a significant increase in the incidence of cop-
ulation and egg laying (Strumwasser et a/., 1969; Audesirk,
1979), as well as increased synthesis of the hormone that
controls egg laying (egg laying hormone; Berry, 1982).
Earlier work has shown that egg laying hormone, a
peptide synthesized and secreted by the neuroendocrine
bag cells, is responsible for triggering egg-laying behavior
(Strumwasser et al.. 1969;Kupfermann, 1970; Arch, 1972;
Dudek et al.. 1980; Stuart et al.. 1980; Chiu and Strum-
wasser, 1981; Blankenship et al., 1983). Although much
is known about the molecular biology of bag-cell peptides
(Chmetal.. 1979; Heller et al.. 1 980; Scheller ?/«/.. 1982;
Mahon and Scheller. 1983) and about the electrophysi-
ological properties of bag cells (Kupfermann and Kandel.
1970; Kaczmarek et al.. 1978, 1982; Kaczmarek and
Strumwasser, 1981), the seasonal regulation of bag-cell
activity and egg laying remains obscure. The general goal
of this and future studies is to gain insight into the mech-
anisms underlying seasonal fluctuations in egg laying be-
havior and in reproductive neuroendocrine function.
LIGHT AND TEMPERATURE AFFECT EGG LAYING
The occurrence of reproductive activity at a particular
time of year suggests the involvement of some environ-
mental timing agent (e.g., ambient temperature, photo-
period, food availability, specific nutritional cue). Previous
studies in Aplysia have shown that warm water can stim-
ulate egg laying, whereas cold temperatures inhibit this
behavior (Berry, 1984; Pinsker and Parsons, 1985). Al-
though the authors interpreted their results to suggest that
changes in the rate of egg laying are solely dependent on
seasonal cycles of temperature, the studies did not test for
effects of other environmental variables, such as photo-
period.
The annual cycle of photoperiod is the most regular
and predictable environmental factor, and is therefore
used by a wide variety of temperate-zone species to time
reproduction to the appropriate season (mammals: Turek
and Campbell, 1979; birds: Rowan, 1926; reptiles: Licht,
1967: insects: Lees, 1966; terrestrial slugs: Sokolove et ai,
1984). A. californica are intertidal organisms, spending
much of their time near the water surface (Audesirk,
1979); thus they would be exposed to annual changes in
day length. A. californica might use photoperiodic infor-
mation, as well as temperature cues, to synchronize re-
production to a particular time of year. The main goal of
this study was to determine whether egg laying behavior
can be influenced by photoperiodic signals.
Materials and Methods
General
Specimens of Aplysia californica (200-300 g) were
captured off the coast of California (approximately 34°N
latitude) by Alacrity Marine Supply, Redondo Beach,
California. At the collection sites, the annual range in
water temperature is from approximately 10 to 20°C (Dan
Stark, Alacrity Marine Supply, pers. comm.), and the an-
nual range in photoperiod is from 1 1 to 15.5 h light/day
(includes 1 h civil twilight). Upon arrival in the laboratory,
animals were maintained in temperature- and light-con-
trolled seawater tanks (475 liters; light intensity at water
surface was 700 lux as measured with a photographic light
meter). Water was recirculated through undergravel filters
within the tanks. Treatment groups (initially, 12 animals
per group; 0-3 animals/group died during the course of
the studies) were maintained in separate tanks, and all
animals were kept in single, perforated plastic buckets (20
cm in diameter) so that each individual could be moni-
tored throughout the studies.
To document the egg laying capability (i.e.. reproduc-
tive maturity) of each animal, atrial gland extract was
injected into the hemolymph of all animals upon arrival
in the laboratory (Nagle et al, 1985). Animals with a ma-
ture reproductive system will lay eggs in response to atrial
gland extract, while immature animals will not lay eggs.
An Aplysia that did not lay eggs spontaneously during
the course of the studies was again treated with atrial gland
extract at the end of each study to assess maturity. Only
those animals that were reproductively mature by the end
of the studies were included in the analysis. Animals were
fed a combination of Romaine lettuce and dried seaweed
(Msubi Nori, Japan Food Corp.) daily. Egg masses were
recorded daily from individual buckets. Because Aplysia
lays eggs at a maximal rate of once per day and does not
consume its own eggs (unpub. obs.), the presence or ab-
sence of an egg mass is an excellent indication of whether
an animal exhibited egg laying behavior on any given day.
The effects of photoperiod on egg laying behavior were
determined as follows. Specimens of Aplysia were col-
lected and shipped to our seawater facilities at four dif-
ferent times of year. Animals that arrived in the early
AUTUMN 1988 (Sept. 22) were all reproductively mature
at the beginning of the study and were divided into four
treatment groups. Aplysia individuals were kept either on
short days (8 h light/day) or on long days ( 16 h light/day);
animals on these two photoperiods were further divided
and maintained either in warm (20°C) or in cold (15°C)
water. Thus the combined effects of photoperiod and water
temperature on egg laying could be investigated. Animals
maintained in cold water rarely, if ever, layed eggs, so we
dropped the cold-water group from the remaining studies.
Aplysia individuals that arrived in the early WINTER
1989 (Jan. 3) and the early SPRING 1989 (Mar. 31) were
reproductively immature at the beginning of the studies;
but they had all reached maturity by the end of the ex-
periments. In these two studies, all animals were main-
tained in warm water and kept either on short or on long
days. Aplysia individuals that arrived in the early SUM-
MER 1989 (June 23) were reproductively mature at the
beginning of the study and were maintained in warm water
and kept either on short or on long days.
The role of the eyes in mediating the effects of photo-
period on egg laying was investigated with specimens of
Aplysia that were brought to the laboratory in the late
SUMMER 1990 (Aug. 7) and maintained in warm water
and 14.25 h light/day (photoperiod in mid-August at 34°N
latitude) for three days. All of these animals were im-
mobilized with MgCl: (injected into hemolymph); half of
them were bilaterally enucleated, and the other half served
as intact controls. Following surgery, animals were further
divided and kept either on short (8 h light/day) or on long
days (16 h light/day), making a total of four treatment
groups.
Analysis of data
Differences in egg laying between treatment groups were
assessed by Chi-square analysis. Values were significantly
different if P< 0.05.
10
N. L. WAYNE AND G. D. BLOCK
100
AUTUMN
I short days, warm
I long days, warm
A. .
short days, cold
long days, cold
B.
C.
6-10
11-15
16-21 1-5 6-10
Days of experiment
11-15 16-21
warm cold
shortday brgday shortday bngday
Figure 1. Percent of.-iplysia individuals laying eggs during the early AUTUMN 1988. Panel A: Animals
were kept in warm (20°C) water and either short (8 h light/day) or long (16 h light/day) days. Data were
averaged (+sem) into 5-day bins. Panel B: Animals were kept in cold (15°C) water and either short (8 h
light/day) or long ( 16 h light/day) days. Data are presented as in panel A. Panel C: Data from the 4 groups
are presented as the percent of animals laying eggs each day, averaged (+sem) over the entire 2 1-day study.
Different letters indicate values are significantly different (at least P < 0.05).
Results
Photoperiodic effects on egg laying
Overall, photoperiod and temperature can both affect
the frequency of egg laying. In the AUTUMN, Aplysia
individuals kept in warm water laid eggs more frequently
than those kept in cold water (Fig. 1). Furthermore, an-
imals maintained on short days and warm water laid eggs
more frequently than those kept on long days and warm
water. However, in the WINTER (Fig. 2) and in the
SPRING (Fig. 3), egg laying frequency overall was sup-
pressed in all groups (even though animals were repro-
ductively mature by the end of the studies; see Materials
and Methods), and there was no apparent effect of pho-
toperiod on egg laying. In the SUMMER (Fig. 4), we once
again observed the emergence of a photoperiodic effect:
Aplysia maintained on short days and warm water laid
eggs more frequently than those kept on long days and
warm water. This photoperiodic response in the summer
was not as robust as that seen during the previous autumn
(compare Figs. Ic and 4b).
Photoperiodic effects in intact vs. eyeless animals
The eyes appear to play a role in transducing photo-
periodic information to the reproductive axis responsible
for regulating egg laying (Fig. 5). Once again, control an-
imals kept in short days and warm water laid eggs more
frequently than those kept on long days and warm water.
WINTER
100
Q.
<
80"
60-
ra 4fj-
20-
short days, warm
long days, warm
B.
50
40 >
"S ~°-
Q) *<
30 ^ £
0) —
20 | |
10 ~ I
0
1-5 6-10 11-15 16-20 21-25 26-30 31-36 short days long days
Days of experiment
Figure 2. Percent of Aplysia individuals laying eggs during the early WINTER 1 989. Animals were kept
in warm (20°C) water and either short (8 h light/day) or long (16 h light/day) days. Panel A: Data were
averaged (+sem) into 5-day bins. Panel B: Data are presented as the percent of animals laying eggs each
day, averaged (+sem) over the entire study. There was no significant difference between the values of the
two groups.
LIGHT AND TEMPERATURE AFFECT EGG LAYING
SPRING
II
iuu -
B short days, warm A-
in
8
80-
long days, warm
TO
c
60 -
_<8
40-
|
20-
~F
3«
^ Wlfa
<c ^><rt -*H-«C *c f^r\ OH oc oc *si
50
B.
^
•
40
II
*< *2.
"
30
< —
If
20
^ ^
10
<a
CO
0
short days long days
Days of experiment
Figure 3. Percent ofAplysia individuals laying eggs during the early SPRING 1989. Data are presented
as in Figure 2.
On the other hand, there was no significant difference in
the frequency of egg laying between the two eyeless groups.
Although the photoperiodic response in the intact control
group was significant, it was not nearly as robust as that
seen in a previous study (see Fig. 1 ).
Discussion
This study provides the first evidence that the repro-
ductive system of Aplysia is responsive to photoperiodic
signals. The results suggest that both photoperiod and
temperature can influence the seasonal rhythm of egg lay-
ing. Specifically, warm temperature is permissive for the
expression of the stimulatory effects of short days. Studies
in another poikilotherm, the lizard Anolis carolinensis,
have also documented that the reproductive response to
stimulatory day lengths is evident in warm, but not cool,
temperatures (Licht, 1967). In addition, recent work in
the edible snail Helix pomalia has shown that egg-laying
behavior is regulated by both photoperiod and tempera-
ture cues (Gomot, 1990). In the wild, the reproductive
activity of Aplysia califomica peaks in late summer-au-
tumn (Strumwasser el al, 1969; Audesirk, 1979; Berry,
1982). At this time of year, water temperature is reaching
a maximum off the coast of California, and day length is
decreasing. Our findings that warm water and short days
stimulate egg laying are therefore consistent with the be-
havior of the animal in its natural environment.
Animals brought to the laboratory in the winter and
spring layed eggs infrequently, if at all, regardless of en-
vironmental treatment. That is, an average of less than
10% of the winter and spring animals laid eggs on any
given day during the course of the two studies — even un-
der stimulatory conditions of short days and warm water.
Although these animals were reproductively immature at
the onset, towards the end of the studies they had reached
maturity and were capable of laying eggs following hor-
monal stimulation (see Materials and Methods). There-
fore, ovotesticular function was most likely not the lim-
iting factor in these studies (however, we do not know
when during the studies animals attained reproductive
maturity).
SUMMER
100
short days, warm
long days, warm
6-10 11-15 16-20 21-25
Days of experiment
2&31 short days long days
Figure 4. Percent of Aplysia individuals laying eggs during the early SUMMER 1989. Data are presented
as in Figure 2. In Panel B, different letters indicate values are significantly different (P < 0.05).
12
N. L. WAYNE AND G. D. BLOCK
100
short days, intact
long days, intact
short days, eyeless
D long days, eyeless
1-5 6-1011-1516-2021-2526-30 1-5 6-1011-1516-2021-2526-30 intact eyeless
Days of experiment short-day long-day short-day long-day
Figure 5. Percent ofAplysia individuals laying eggs during the late SUMMER 1 990. All animals were
maintained in warm (20°C) water. Panel A: Intact, control animals were kept on short (8 h light/day) or
long (16 h light/day) days. Data were averaged (+sem) into 5-day bins. Panel B: Bilaterally enucleated
animals were kept on short (8 h light/day) or long (16 h light/day) days. Data are represented as in panel
A. Panel C: Data from the 4 groups are presented as the percent of animals laying eggs each day, averaged
(+sem) over the entire 30-day study. The letters a and b indicate values that are significantly different (P
< 0.05). 'Indicates that values approached significant difference compared to that of the intact, long-day
control group (P < 0. 10).
A common phenomenon among some seasonally
breeding vertebrates is a spontaneous shutdown of the
reproductive system during the non-breeding season (liz-
ard: Cueller and Cueller, 1977; birds: Hamner, 1967;
Robinson and Follett, 1982; mammal: Robinson and
Karsch, 1984). During this period of reproductive qui-
escence, previously inductive photoperiodic cues no longer
stimulate reproductive activity. This period of insensitivity
to stimulatory photoperiod (commonly labelled 'photo-
refractoriness') is an endogenous process and can be 'bro-
ken' by exposing the animal to a bout of inhibitory pho-
toperiod, followed by a stimulatory day length (Jackson
el ai, 1988). In Aplysia. we have shown that previously
stimulatory environmental cues (warm water, short days)
did not stimulate spontaneous egg laying during the non-
breeding season in winter and spring. Aplysia may there-
fore behave like many other seasonal breeders and become
refractory to stimulatory signals. If this is so, then pre-
treatment with long days and cold temperatures should
be able to break refractoriness to stimulatory short days
and warm temperatures.
But mechanisms other than an endogenous refracto-
riness to an environmental signal might equally well un-
derlie the cessation of spontaneous egg laying by Aplysia
during winter and spring. For instance, one or more key
components of the reproductive neural axis may be de-
velopmentally immature during the winter and spring
(even though the reproductive tract can mature in the
laboratory). Alternatively, some environmental cue (e.g.,
food or other nutritional item necessary for high levels of
spontaneous egg laying) may be missing during that time
of year.
Further, our results suggest that the eyes play a role in
mediating photoperiodic information to the reproductive
axis responsible for regulating egg laying behavior. Spe-
cifically, photoperiod had no effect on egg laying in those
animals that were bilaterally enucleated. The eyes of
Aplysia contain both photoreceptors and a circadian
pacemaker (Jacklet, 1969; Eskin, 1971). The circadian
system is involved in the neural pathway mediating pho-
toperiodic responses in most animals investigated (Follett
and Sharp, 1969; Elliott, 1976; Almeida and Lincoln,
1982). Furthermore, both ocular and extraocular photo-
receptors mediate photoperiodic responses in a variety of
species (Reiter, 1969; Follett et a/.. 1975; Legan and
Karsch, 1983; Foster and Follett, 1985). In Aplysia, pho-
toreceptors are found not only in the eye, but also in
structures as diverse as the abdominal ganglion (Andresen
and Brown, 1982), the cerebral ganglion (Block and Smith,
1973), the rhinophores (Jacklet, 1980), the oral veil (Cook
et ul. . 1 99 1 ). and the siphon ( Lukowiak and Jacklet, 1972).
Therefore, the relative roles of the ocular photoreceptors
and ocular pacemakers in mediating the effects of pho-
toperiod on egg laying are not clear. For instance, both
the ocular photoreceptors and ocular pacemakers may be
playing a role in photoperiodic time measurement. Al-
ternatively, extraocular photoreceptors may be transmit-
ting light signals to the ocular circadian pacemaker, which
then sends its signals to the next step in the photoperiodic
response system.
Nevertheless, we must stress that our results are difficult
to interpret, because the photoperiodic response in the
intact controls in the last experiment was weak compared
to that seen in the first experiment (compare Fig. 1 with
LIGHT AND TEMPERATURE AFFECT EGG LAYING
13
Fig. 5). An intriguing mystery arising from these studies
is the source of the variability in the photoperiodic re-
sponse from one year to the next. Because we work with
animals captured in the wild, we have no control over the
environmental history of the animal. That is, we cannot
control for variations in microhabitat (i.e., local and year-
to-year variability in water temperature, food availability,
sexual experience). Large, year-to-year fluctuations in the
availability of the algae that Aplysia feed upon were ob-
served during the course of these experiments: algae were
abundant in the summer to early autumn of 1988, but
scarce in the summer to early autumn of 1989 and 1990
(Dan Stark, Alacrity Marine Supply, pers. comm.); and
these fluctuations were associated with similar changes in
the robustness of the photoperiodic response. In addition
to photoperiodic and temperature signals, there are other
environmental variables (e.g., food or nutritional cues)
that might affect spontaneous egg laying. All of these en-
vironmental cues may act in combination such that one
variable alters the effectiveness of the other variables on
the frequency of egg laying. For instance, the photoperi-
odic response during the breeding season may be more
robust in those animals with a high level of nutrition; low
nutrition may weaken the photoperiodic response. Food
availability has pronounced effects on the photoperiodic
diapause response in some insect species (Saunders, 1979).
Future work investigating the role of food cues may pro-
vide some information on its importance in the expression
of a robust photoperiodic response in Aplysia.
Acknowledgments
This work was supported by NIH grants NS-08725 to
NLW and NS- 15264 to GDB.
Literature Cited
Andresen, M. C., and A. M. Brown. 1982. Cellular basis of the pho-
toresponse of an extraretinal photoreceptor. Experienlia 38: 1001-
1006.
Almeida, O. F. X., and G. A. Lincoln. 1982. Photoperiodic regulation
of reproductive activity in the ram: evidence for the involvement of
circadian rhythms in melatonin and prolactin secretion. Biol. Reprod.
27: 1062-1075.
Arch, S. 1972. Biosynthesis of the egg-laying hormone (ELH) in the
bag cell neurons of Aplysia californica. J. Gen. Physiol. 60: 102-1 19.
Audesirk, T. E. 1979. A field study of growth and reproduction in
Aplysia californica. Biol. Bull. 157: 407-421.
Berry, R. \V. 1982. Seasonal modulation of synthesis of the neuro-
secretory egg-laying hormone of Aplysia. J Neurobiol. 13: 327-335.
Berry, R. W. 1984. Environmental temperature modulates the rate of
synthesis of egg-laying hormone in Aplysia. J. Comp. Physiol. B 154:
545-548.
Blankenship, J. E., M. K. Rock, L. C. Robbins, C. A. Livingston, and
H. K. Lehman. 1983. Aspects of copulatory behavior and peptide
control of egg laying in Aplysia. Fed. Proc. 42: 96-100.
Block, G. D., and J. T. Smith. 1973. Cerebral photoreceptors in Aplysia.
Comp. Biochem. Physiol. 46a: 115-121.
Chiu, A. Y., and F. Strumwasser. 1981. An immunohistochemical study
of the neuropeptidergic bag cells of Aplysia. J. Neurosci. 1:81 2-826.
Chiu, A. Y., M. W. Hunkapiller, E. Heller, D. K. Stuart, L. E. Hood,
and F. Strumwasser. 1979. Purification and primary structure of
the neuropeptide egg-laying hormone of Aplysia californica. Proc.
Nail. Acad. Sci. 76: 6656-6660.
Cook, D. G., M. Stopfer, and T. J. Carew. 1991. Identification of a
reinforcement pathway necessary for operant conditioning of head
waving in Aplysia californica. Behav. Neural Biol. 55: 313-337.
Cueller, H. S., and O. Cueller. 1977. Refractoriness in female lizard
reproduction: a probable circannuai clock. Science 197: 495-497.
Dudek, F. E., G. Weir, J. Acosta-Urquidi, and S. S. Tobe. 1980. A
secretion from neuroendocrine bag cells evokes egg release in vitro
from ovotestis of Aplysia californica. Gen. Comp. Endocrinol. 40:
241-244.
Elliott, J. A. 1976. Circadian rhythms and photopenodic time mea-
surement in mammals. Fed. Proc. 35: 2339-2346.
Eskin, A. 1971. Properties of the Aplysia visual system: in vitro en-
trainment of the circadian rhythm and centrifugal regulation of the
eye. Z r?/. Physiol. 74: 353-371.
Follett, B. K., and P. J. Sharp. 1969. Circadian rhythmicity in pho-
toperiodically induced gonadotropin release and gonadal growth in
the quail. Nature 223: 968-97 1 .
Follett, B. K., D. T. Davies, and V. Magee. 1975. The rate of testicular
development in Japanese quail (Coturnix cotumix japonica) following
stimulation of the extra-retinal photoreceptor. Experienlia 31: 48-
49.
Foster, R. G., and B. K. Follett. 1985. The involvement of rhodopsin-
like photopigment in the photoperiodic response of the Japanese
quail. / Comp. Physiol. A 157: 519-528.
Gomot, A. 1990. Photopenod and temperature interaction in the de-
termination of reproduction of the edible snail. Helix pomatia. J.
Reprod. Pert. 90: 581-585.
Hamner, W. M. 1967. The photorefractory period of the house finch.
Ecology 49: 211-227.
Heller, E., L. K. Kaczmarek, M. W. Hunkapiller, L. E. Hood, and F.
Strumwasser. 1980. Purification and primary structure of two neu-
roactive peptides that cause bag cell afterdischarge and egg-laying in
Aplysia. Neurobiology 77: 2328-2332.
Jacklet, J. W. 1969. Circadian rhythm of optic nerve impulses recorded
in darkness from isolated eye of Aplysia. Science 164: 562-563.
Jacklet, J. W. 1980. Light sensitivity of the rhinophores and eyes of
Aplysia. J. Comp. Physiol. 136: 257-262.
Jackson, G. L., M. Gibson, and D. Kuehl. 1988. Photoperiodic dis-
ruption of photorefractoriness in the ewe. Biol. Reprod 38: 127-134.
Kaczmarek, L. K., and F. Strumwasser. 1981. The expression of long
lasting afterdischarge by isolated Aplysia bag cell neurons. / Neurosci.
1: 626-634.
Kaczmarek, L. K., K. Jennings, and F. Strumwasser. 1978. Neuro-
transmitter modulation, phosphodiesterase inhibitor effects, and cyclic
AMP correlates of afterdischarge in peptidergic neuntes. Proc. Natl.
Acad. Sci. 75: 5200-5204.
Kaczmarek, L. K., K. R. Jennings, and F. Strumwasser. 1982. An early
sodium and late calcium phase in the afterdischarge of peptide-se-
creting neurons in Aplysia. Brain Res. 238: 105-1 15.
Kupfermann, I. 1970. Stimulation of egg laying by extracts of neu-
roendocrine cells (bag cells) of abdominal ganglion of Aplysia. J
Neurophysiol. 33: 877-881.
Kupfermann, L, and E. R. Kandel. 1970. Electrophysiological properties
and functional interconnections of two symmetrical clusters (bag cells)
in abdominal ganglion of Aplysia. J Neurophysiol. 33: 865-876.
14
N. L. WAYNE AND G. D. BLOCK
Legan, S. J., and F. J. Karsch. 1983. Importance of retinal photore-
ceptors to the photoperiodic control of seasonal breeding in the ewe.
Biol. Reprod. 29: 316-325.
Lees, A. D. 1966. Photoperiodic timing mechanisms in insects. Nature
210: 986-989.
Licht, P. 1967. Environmental control of annual testicular cycles in the
lizard Anolis carolinensis. I. Interaction of light and temperature in the
initiation of testicular recrudescence. / Exp. Zoo/. 165: 505-516.
Lukowiak, K., and J. W. Jacklet. 1972. Habituation and dishabituation:
interactions between peripheral and central nervous systems in Aply-
sia. Science 178: 1306-1308.
Mahon, A. C., and R. H. Scheller. 1983. The molecular basis of neu-
roendocnne fixed action pattern: egg laying in Aplysia. Cold Spring
Harbor Symp. Quant. Biol. 48: 405-412.
Nagle, G. T., S. D. Painter, K. L. Kelner, and J. E. Blankenship.
1985. Atrial gland cells synthesize a family of peptides that can
induce egg laying in Aplysia. J Comp Physiol B 156: 43-55.
Pinsker, H. M., and D. W. Parsons. 1985. Temperature dependence
of egg laying in Aplysia brasiliana and A. califomica. J Comp. Physiol.
B 156: 2 1-27.
Reiter, R. J. 1969. Pineal function in long term blinded male and
female golden hamsters. Gen. Comp. Endocrinol. 12: 460-468.
Robinson, J. E., and B. K. Follett. 1982. Photoperiodism in Japanese
quail: the termination of seasonal breeding by photorefractoriness.
Proc. R. Soc. London B 215: 95-1 16.
Robinson, J. E., and F. J. Karsch. 1984. Refractoriness to inductive
day lengths terminates the breeding season of the Suffolk ewe. Biol.
Reprod. 31: 656-663.
Rowan, W. 1926. On photoperiodism, reproductive periodicity, and
the annual migrations of birds and certain fishes. Proc Boston Soc.
Natl. Hist. 38: 147-189.
Saunders, D. S. 1979. Insect Clocks. Pergamon Press, New York. Pp.
102-107.
Scheller, R. H., J. F. Jackson, L. B. McAllister, J. H. Schwartz, E. R.
Kandel, and R. Axel. 1982. A family of genes that codes for ELH,
a neuropeptide eliciting a stereotyped pattern of behavior in Aplvsia.
Cell 28: 707-7 19.
Sokolove, P. G., E. J. McCrone, J. van Minnen, and W. C. Duncan.
1984. Reproductive endocrinology and photoperiodism in a ter-
restrial slug. Pp. 189-203 in Photoperiodic Regulation of Insect and
Molluscan Hormones. Pitman, London.
Strumwasser, F., J. W. Jacklet, and R. B. Alvarez. 1969. A seasonal
rhythm in the neural extract induction of behavioral egg-laying in
Aplysia. Comp. Biochem. Physiol. 29: 197-206.
Stuart, D. K., A. Y. Chiu, and F. Strumwasser. 1980. Neurosecre-
tion of egg-laying hormone and other peptides from electrically
active bag cell neurons of Aplysia. J Neurophysiol. 43: 488-
498.
Turek, F. W., and C. S. Campbell. 1979. Photoperiodic regulation of
neuroendocrine-gonadal activity. Biol. Reprod. 20: 32-50.
Reference: Biol. Bull. 182: 15-30. (February, 1992)
The Development and Larval Form
of an Echinothurioid Echinoid,
Asthenosoma ijimai, Revisited
S. AMEMIYA1 AND R. B. EMLET2*
lMisaki Marine Biological Station, Miura-shi, Kanagawa 238-02, Japan and -Department of
Biological Sciences, University of Southern California, Los Angeles, California 90089-0371
Abstract. The modified development from cleavage to
late larval form of the echinothurioid echinoid, Asthe-
nosoma ijimai, was re-examined using light microscopy
and scanning electron microscopy of whole and sectioned
stages. Although an original study (Amemiya and Tsu-
chiya, 1979) reported direct development without evi-
dence of a pluteus larva, we found that the unusual de-
velopment can be interpreted as a topologically reflected,
reduced pluteus, with vestigial larval arms and a greatly
reduced larval skeleton. This developmental pattern pro-
duces the third and most reduced pluteus form known
among the six echinoid lineages with modified develop-
ment that have been studied thus far. Features such as an
equal fourth cleavage, extrusion of yolk into the blastocoel,
and the presence of large numbers of cells within the blas-
tocoel are convergent with traits reported for other species
with modified development. Coelom formation is clearly
modified from that of species with feeding larval devel-
opment, but notably the hydrocoel begins to develop po-
dial buds prior to separation from the archenteron.
Echinothurioid sea urchins are considered to be the most
primitive living euechinoids, and in A. ijimai the timing
of mesenchyme cell ingression and the formation of epi-
neural folds were similar to these features in other eue-
chinoids. Indentation of the juvenile oral surface relatively
late in larval development raises the possibility that the
amniotic invagination (vestibule), common in all other
euechinoids, may be a trait incorporated into the devel-
opment of echinoids at the time of origin of the echino-
thurioids. The structural comparisons reported here show
Received 31 July 1991; accepted 25 November 1991.
* Order of authorship was determined alphabetically.
a need for further detailed morphological studies of de-
velopmental modifications in other echinoid species.
Introduction
Most sea urchins species (ca. 66%) develop through a
feeding larval stage, the echinopluteus, for several to many
weeks before metamorphosing into juvenile echinoids (see
review by Emlet et ai, 1987). At least 14 times among
living taxa, however, the feeding larval stage has been lost,
and these species undergo a modified and abbreviated
development before juvenile sea urchins are formed (Em-
let, 1990; see also Strathmann, 1978; Raff, 1987). At pres-
ent, about ten species from six of the lineages with mod-
ified development have been investigated, and some de-
scription of their embryonic and larval development is
available. Descriptive and analytical research has been
conducted on the following: two cidaroids (Phyllacanthus
imperialis, Olson et ai. 1988; P. pan'ispinus, Mortensen,
1921; Parks et ai, 1989); two echinothurioids (Astheno-
soma ijimai, Amemiya and Tsuchiya, 1979; A. sp., Uehara
and Amemiya, unpub. obs.); one temnopleuroid (Ho-
lopneustes purpiirescens, V. Morris, Univ. Sydney, un-
pub.); one echinometrid (Heliocidaris erythrogramma,
Mortensen, 1921; Williams and Anderson, 1975; Parks
et ai, 1988); two clypeasteroids (Peronella japonica, Mor-
tensen, 1921; Okazaki and Dan, 1954; Okazaki, 1975; P.
rubra. Amemiya and Emlet, unpub. obs.); and two
brooding spatangoids (Abatus agassizi, Larrain. 1973; A.
cordatus, Schatt, 1985, 1988). Many of the other echinoid
lineages with modified development occur in deeper seas
or antarctic seas and are difficult to collect for study (e.g..
lineages oftemnopleuroids, holasteroids and other lineages
of cidaroids and spatangoids, <.•./.", Mortensen, 1936; Fell,
1976).
15
16
S. AMEMIYA AND R. B. EMLET
The most recent studies have focused on species with
very highly modified development, often referred to as
direct development. These studies have examined heter-
ochrony (changes in relative timing) of developmental
events, modifications of cleavage patterns, the resultant
cell lineages, and cell movements (e.g.. Parks et al.. 1988,
1 989; Wray and Raff, 1989, 1990; Henry and Raff, 1990).
Additional immunofluorescence studies have drawn in-
ferences about gene expression from specific markers for
gene products [e.g., the monoclonal antibody B2C2 to
mesenchyme-derived antigens or antibodies to seroton-
ergic neurons (above citations, Bisgrove and Raff, 1989)].
Due to the extreme degree of developmental modifica-
tions, these studies usually emphasize how different the
morphogenetic patterns are from those of species that de-
velop through a pluteus larva (e.g., Heliocidaris erythro-
gramnia, Wray and Raff, 1990). With the exception of
the above mentioned studies on Ahatus cordatits, Helio-
cidaris erythrogramma, and Peronellajaponica, either no
information or only limited information is available on
the internal morphological aspects of development of
species with modified development. Both the extreme
modification, and a lack of detailed morphological infor-
mation, limit our understanding of how these develop-
mental modifications may have evolved.
This report — an extension of an earlier one (Amemiya
and Tsuchiya, 1979) — describes selected features of mor-
phogenesis in the echinothurioid echinoid Asthenosoma
ijimai, from cleavage through late larval development.
Comparisons are also made with unmodified pluteal de-
velopment, as well as with the modified developmental
patterns occurring in other species. The echinothurioids
are a particularly important group to study: first, because
all of them seem to have modified development (reviewed
in Emlet et al., 1987); and second, because they are con-
sidered to be the earliest living branch of the euechinoid
lineage, and thus the second oldest lineage of echinoids
after the cidaroids (Smith, 1984). Because the cidaroids
and euechinoids differ in many developmental features
(Emlet, 1988), we should inquire whether the develop-
ment of the echinothurioids shows greater affinity with
other euechinoids, or with the more primitive cidaroids.
The observations presented here, together with the com-
parisons with other species, point up the need for addi-
tional morphological studies of developmental modifi-
cations in other echinoid species.
Materials and Methods
Adults and larvae
Adults of Asthenosoma ijimai Yoshiwara were collected
at a depth of 20 m off Misaki Marine Biological Station
in Sagami Bay, Japan. Adult specimens were dissected,
and fully matured gametes were obtained. Eggs were
washed twice in filtered seawater (0.22 /urn) and fertilized
by mixing with a small amount of undiluted sperm. Fer-
tilized eggs were washed three times in filtered seawater,
and cultured in unstirred, one-liter glass beakers at 20°C
or at room temperature (25-28°C). The stages and times
for sectioned material presented here are from cultures at
room temperature. No food was added to the larval cul-
tures. At various times after fertilization, living larvae were
photographed under a dissecting microscope, and aliquots
were fixed for examination by light or scanning electron
microscopy (SEM).
Preparation of sectioned and stained material
Larvae were fixed for 1 h in seawater containing 4% or
10% formalin at room temperature and preserved in 70%
EtOH. Preserved specimens were dehydrated through
graded ethanol series and embedded in Spurr embedding
media (Polysciences, Inc). Sections, 5-8 f*m thick, were
stained with Richardson's stain (1% Azure II in distilled
water combined with 1%. methylene blue in 1% sodium
borate, Richardson et al., 1960). The serial sections of
larvae embedded in epoxy resin were traced with camera
lucida and digitized so that 3-D images of the sections
could be constructed (PC3D program, Jandel Scien-
tific, Inc.).
Immunofluorescence and H33258 staining
Immunofluorescence staining with skeletogenic mes-
enchyme specific monoclonal antibody B2C2 was con-
ducted according to the methods of Parks et al. (1988).
Embryos, larvae, and juveniles were fixed for 50 min in
seawater containing 4% formalin, washed in artificial sea-
water, dehydrated in a graded ethanol series, embedded
in polyester wax (BDH, Ltd), and sectioned at a thickness
of 5 ^m. The rehydrated sections were washed with phos-
phate-buffered saline containing 0.05% Tween 20 (PBS-
TW20) and incubated with culture fluid containing the
monoclonal antibody B2C2, diluted 1 :20 in PBS-TW20,
for 40 min at room temperature in a humidified chamber.
The slides were washed in PBS-TW20, incubated with
FITC-conjugated, goat anti-mouse, IgG antibodies (di-
luted 1:200 in PBS-TW20) for 40 min, and rinsed again.
For detection of cell nuclei, some sections were incubated
with H33258 (Hechst, Inc.) at a concentration of 0.5 /ug/
ml PBS for 10 min instead of, or after, treatment in pri-
mary and secondary antisera. Fluorescence was observed
and photographed with a Nikon fluorescence microscope.
Scanning Electron Microscopy (SEM)
Specimens were fixed and preserved as indicated above,
or they were fixed for 1 h in a mixture of 2% gluteraldehyde
(Taab Lab.) and 1% osmium tetroxide (OsO4, Taab Lab.)
ECHINOTHURIOID DEVELOPMENT, REVISITED
17
Figure 1. Early embryonic stages ofAsthenosoma ijimai. a. Sixteen-cell stage embryos randomly oriented,
show that all cells are approximately the same diameter after the fourth cleavage. Scale bar. 1 mm. b. SEM
of a 21.5-h, lobate blastula. Arrowheads mark pits in ectoderm. Scale bar, 1 mm. c. Close-up SEM of
ectodermal pit marked by right arrow in b. Scale bar, 25 nm. d. Section of a 2 1 ,5-h blastula shows yolky
cytoplasm in the blastocoel and ectodermal pits (arrowheads). Same scale as b.
in 0.45 M sodium acetate buffer (pH 6.4) at room tem-
perature (Harris and Shaw, 1984) and preserved in 70%
EtOH at 4°C. The preserved specimens were dehydrated
through a graded ethanol series, dried at the critical point
(Hitachi HCP-1 drier) with liquid CO: as a transitional
fluid, and sputter-coated with gold (Eiko IB-3 ion coater).
Observations were made with a Hitachi HHS-2R SEM.
To examine the inside of larvae with SEM, specimens
were embedded in polyester wax and sectioned by micro-
tome to expose a particular cross section. These specimens
were incubated in absolute ethanol at 40°C for 12 h to
remove wax (Armstrong and Parenti, 1973) and then
subjected to critical point drying as described above.
Clearing lan>ae
Live larvae ofAsthenosoma are opaque orange-yellow
(Amemiya and Tsuchiya, 1979), and it was impossible to
see internal structures in these or in fixed, preserved spec-
imens. However, larvae could be rendered translucent by
clearing with solutions of benzyl benzoate and benzyl al-
cohol mixed in ratios of 2:1, 1:1, or 1:2, depending on
the desired refractive index. Fixed larvae were first de-
hydrated to 100% EtOH, then transferred into the clearing
solution where the remaining EtOH was allowed to evap-
orate. Upon clearing, larval dimensions remained the
same, and no osmotic effects were discerned. To search
for calcareous deposits, cleared larvae were observed under
crossed-polarized light.
Results
Obsen'ations on soft-tissue development
Eggs, cleavage, external aspects of larvae, and meta-
morphosis have been described by Amemiya and Tsu-
chiya (1979). The fourth cleavage of embryos ofAsthen-
osoma ijimai was almost equal, giving rise to 16 blasto-
meres of similar size. Figure la shows that there was some
18
S. AMEMIYA AND R. B. EMLET
•^•'^W-vi '- .
1
^£%l^*t J
•>,. " •« * *£>«**,1GsU5l ^ "<V-3,
- r <-?"•» -~,J*G!.-<&*?'fj,'** **' t
rf^^^^^l*
IU^>**Jfei*^il3
&nt^>«wl
» JSJ^lf'1*! ^^?«
Figure 2. Later embryonic stages of Asthenosoma ijiinai. All embryos and sections are oriented with
the animal pole up. a. Light micrograph of a live, 25.5-h-old, early gastrula. Same scale as b. b. Section
through animal-vegetal axis of flattened gastrula (25.5 h), showing blastocoel filled with yolky cytoplasm.
Scale bar, 1 mm. c. Lateral wall of gastrula (25.5 h), shows yolky cytoplasm exocylosing from basal ends of
ectodermal cells. Same scale as d. d. Vegetal wall of gastrula (25.5 h), showing exocytosis of cytoplasm and
ingression of cells. Scale bar, 200 urn. e. Section of a 25.5-h gastrula shows fluorescently staining nuclei
(H33258 fluorescent dye) of ectodermal and probable mesenchyme cells. Same scale as f. f. Section of 35-
h gastrula, with fluorescently stained nuclei. Scale bar, 200 Mm. g. Section of 51.5-h embryo, stained with
B2C2 antibody, arrow shows first occurrence of expression of MSP- 130 glycoprotein associated with blas-
tocoelic cells. Scale bar, 50 ^m. h, i. Section of an 88.5-h larva, doubly stained with B2C2 antibody (h) and
H33258 fluorescent dye (i) shows not all blastocoelic cells express MSP- 130. The concentrated clusters of
nuclei in (i) are epithelia of the archenteron (A) and a coelomic compartment (C). Scale bar, 200 ^m.
variation in size of the blastomeres in 16-cell embryos,
but there was no evidence of micromeres at the vegetal
pole. The absence of an unequal, fourth cleavage parallels
the cleavage patterns of other echinoid species large yolky
eggs and modified development (Williams and Anderson,
1975; Raff, 1987; Parks et til., 1989).
Like other echinoderm species with yolky eggs, embryos
of Asthenosoma ijimai formed wrinkled blastulae (Ame-
miya and Tsuchiya, 1979; Parks et at., 1989). This stage
was followed by egression and loss of wrinkles and led to
a lobate blastula (Fig. Ib). At this stage, small, cylindrical
pits were present on the external surface of the embryo
and passed into the ectodermal layer (Fig. Ib, c, d). Serial
sections showed that some of these pits terminated blindly
within the ectoderm, while others passed through the ec-
toderm to another external opening. Several of these pits
or passages coincided with large indentations in the em-
bryo's surface and, therefore, may have been the remnants
of the wrinkled indentations. Parks et at. (1989) reported
pits in the yolky embryo of the cidaroid Phyllacanthus
pan'ispimts, though they did not see any association of
pits with egression tracks, where wrinkles diminished.
Further work is necessary to determine whether the pits
in the two species arise by similar mechanisms.
ECHINOTHURIOID DEVELOPMENT. REVISITED
19
Figure 3. A 5 1 .5-h larvae of Asthenosoma ijimai, all with anterior end up. a. Light micrograph of a live
larva with dorsal swelling on the right. Scale bar, 1 mm. b. Medial sagittal section through late gastrula. The
tip of archenteron (A) has curved toward the ventral surface. Same scale as a. c. Frontal section shows a
small outpocketing on the left side of the archenteron (A) that extends dorsally in other sections. Scale bar.
200 /im.
Sections of this 21.5 h stage showed anucleate, yolky
cytoplasm being released into the blastocoel from the basal
ends of most ectodermal cells (Fig. Id, see also Fig. 2c,
d). Sections also revealed that one indented surface was
extruding considerably more material than other surfaces
into the blastocoel (Fig. Id). The fluorescent stain,
H33258, revealed considerable numbers of nuclei in the
blastocoel, indicating mesenchyme-like cell ingression at
the onset of gastrulation.
Gastrulation had begun by 25.5 h after fertilization,
and embryos were compressed along the animal-vegetal
axis, with a large indentation at the vegetal pole (Fig. 2a).
At this stage, the blastocoel was filled with yolky cytoplasm
(Fig. 2b, c, d). Counts of fluorescently stained nuclei
showed a mean of 98 cells per 5-^m section (n = 3 sections,
S.D. = 11) and were scattered among the yolky cytoplasm
in the blastocoel (Fig. 2e). The number of staining nuclei,
and thus the number of cells, increased to a mean of 375
per section (n = 3 sections, S.D. = 9.6) in the mid-gastrula
stage at 35 h (Fig. 2f). The source of these additional
blastocoelic cells is either ingression from the vegetal pole
(Fig. 2d) or cell division. Fluorescent staining also showed
that cell division in the ectoderm was continuing because
the number of nuclei in the ectoderm increased between
25.5 and 35 h (Fig. 2e, f).
A positive reaction of B2C2 antibody with blastocoelic
cells, indicating expression of MSP- 1 30 glycoprotein, was
found first at 51.5 h post fertilization (Fig. 2g). Later ob-
servations on mesenchyme cells associated with skeleton
in larvae of Asthenosoma ijimai showed that these cells
reacted with B2C2, suggesting that MSP- 130 is expressed
by the skeletogenic cells in this species just as in other
echinoids. By 88.5 h post fertilization, after skeletogenesis
had begun, sections labeled with both nuclear stain
(H33258) and B2C2 antibody revealed that only a fraction
of blastocoelic cells were skeletogenic (Fig. 2h, i).
In the present paper, the identities of the dorsal and
ventral surfaces are reversed from those described in the
initial paper on development of Asthenosoma ijimai
(Amemiya and Tsuchiya, 1979). By 35 h post fertilization,
the embryos elongated along the animal-vegetal axis, with
the blastopore located off center, toward the ventral sur-
face. At 5 1 .5 h, the late gastrulae had swollen dorsal sides
(Fig. 3a). Internally, the archenteron had grown over half
the length of the embryo, and the apical (anterior) end
curved toward the ventral side of the embryo (Fig. 3b).
In addition to the large curved tip of the archenteron,
another small outpocketing was forming on the left side
of the archenteron and was growing dorsally (Fig. 3c).
Serial sections of stages at 51.5, 56.5, and 63 h post-
fertilization showed progressive changes in the develop-
ment of the archenteron and coelomic pouches (Fig. 4).
When the archenteron reached its full length, one or two
slender, epithelial projections on the left side and an-
20
S. AMEMIYA AND R. B. EMLET
DORSAL VIEW
Ls
LEFT LATERAL VIEW
/ ^/\h d*/
Figure 4. Three-dimensional reconstructions of the archenteron and coelomic pouches from serial sections
of 51. 5-, 56.5-, and 63-h larvae ofAsthenosoma ijiinai. In each column the same larva is shown in approximate
dorsal view and approximate left lateral view. The animal pole, corresponding to the anterior end, is toward
the top of the figure: the vegetal pole, corresponding to the posterior end, is toward the bottom of the figure.
Line segments are tracings of the inner surface of the archenteron and coelomic pouches (black). The outer
surface of the larval ectoderm (gray) is included in a. c, d and f for reference. Small arrows, left lateral and
dorsal projections from archenteron; large arrow, tip of archenteron; Rs, right somatocoel; Ls. left somatocoel;
H, hydrocoel; Pb, podial bud of hydrocoel. See text for explanation, a., b. 51.5 h. c.. d. 56.5 h. e., f. 63 h.
Scale bar. 1 mm.
terodorsal surface of the archenteron grew dorsally (small
arrows, Fig. 4a, b). Of the three 51.5-h specimens that
were serially sectioned, two showed both projections (Fig.
4a, b), and one showed a left lateral projection only (not
figured). With further development, the much larger ven-
tral tip of the archenteron bent and extended ventrally
without contacting the blastocoel wall (large arrow. Fig.
4b, c). Between 51.5 and 56.5 h, the archenteron under-
went torsion, twisting approximately 90° counterclock-
wise, when viewed from the animal pole. This twist re-
oriented the tip of the archenteron toward the left side of
the larva, and the slender projections toward the right
side of the larva (Fig. 4b, c). One of the two slender pro-
jections extended laterally and posteriorly (toward the
vegetal end) to form the right somatocoel (Fig. 4c-f).
Comparisons among serially sectioned larvae suggest that
either of the slender projections could form the right so-
matocoel, and the other slender projection apparently did
not continue to grow. Subsequent to 56.5 h, the tip of the
archenteron grew a projection dorsally and posteriorly,
which formed the left somatocoel (Fig. 4e, f). By 63 h.
while still attached to the main body of the archenteron,
the tip of the archenteron began to develop into the left
hydrocoel with buds that became the coelomic lining of
the five, primary podia (Fig. 4e, f; Fig. 5c, d).
Externally, changes between the 56.5 and 63 h stages
produced four large and rounded lobes that grew into
projections called "para-arms" by Amemiya and Tsuchiya
(1979). One pair of these bilaterally symmetrical projec-
tions is located dorsally and laterally relative to the blas-
topore and projects posteriorly, away from the animal
pole. The second pair is located on the dorsal surface just
anterior to the other pair and also projects dorsally (Fig.
5a). The surfaces of the larvae were uniformly ciliated
(Fig. 5b). No developing stages, even in the region of the
para-arms, showed cilia collected into discrete rows such
as found in the ciliated bands of pluteus larvae. No dorsal
hydropore was present despite internal development of
somatocoels and the left hydrocoel.
By 75.5 h post-fertilization, the five bulges of the pri-
mary podia were externally visible and arranged in a circle
on the left lateral surface (Fig. 6a). Sections of this stage
ECHINOTHURIOID DEVELOPMENT, REVISITED
21
V?**
Figure 5. A 63-h larvae of Asthenosoma ijimai. all oriented with the anterior end toward the top of the
figure, a. SEM of a whole larva, right ventral view (dorsal side is on the left) shows two right para-arms
(arrows) and the blastopore (Bp). Scale bar. 0.5 mm b. Close-up of uniformly ciliated epidermis. Scale bar,
25 nm. c. Frontal section shows the leftward oriented archenteron with coelomic components of podia near
its tip. The hydrocoel (H) is developing before being separated from the archenteron (A). (This larva was
damaged during embedding, but a clear interpretation of sections was still possible.) Same scale as a. d.
Higher magnification of a more dorsally located frontal section from same larva as c. Hydrocoelic components
of podial buds (Pb) are present. Ls. left somatocoel. Scale bar. 200 ^m.
22
S. AMEMIYA AND R. B. EMLET
Figure 6. A 75.5-h larvae of Asthenosoma ijimai. a. Light micrograph of live specimen shows five
primary podia just beginning to form on left side of larva. Scale bar, 0.5 mm. b. Frontal section with
continued hydrocoelic (and podial) development. The hydrocoel (H) is almost completely separated from
the archenteron (A). Same scale as a. c. Detail of hydrocoel (H) with parts of two podial extensions from a
different section of the same larva as b. Scale bar. 100 ^m.
showed the hydrocoelic compartments with thickened
epithelia beneath the podial swelling of the ectoderm (Fig.
6c). Serial sections revealed that the connection between
the hydrocoel and archenteron was greatly reduced in one
larva (Fig. 6b) and completely severed in a second larva
examined. All coelomic and archenteric cavities contained
stained materials that appeared to be yolky cytoplasm
and some cells (Fig. 6b, c).
By 101 h post fertilization, primary podia elongated to
0.2 mm length (Fig. 7b, c, d). Sections of the juvenile oral
surface showed folds of ectodermal tissue lying between
the five primary podia (Fig. 7c, d). These folds were evi-
dently epineural folds that were growing over the juvenile
oral surface to form the epineural sinus (von Ubisch, 1913;
Hyman, 1955; Emlet, 1988). SEM observations of the
external surface of the developing juvenile oral region
confirm that these epineural folds were spreading toward
the oral center (Fig. 7f-h).
Coincident with the lengthening of the primary podia
and development of the epineural folds, the oral surface
sank to become indented in the surface of the developing
larva. This indentation was notable in live specimens
viewed from the side at 101 h (Fig. 7b), as well as in sec-
tioned material (Fig. 7c) and in specimens fixed for SEM
(Fig. 7f ). Though the developing juvenile oral surface was
never deeply enclosed as occurs within the amniotic in-
vagination (or vestibule) of the euechinoids, the oral sur-
face was further sunken in living larvae nine days post-
fertilization (Fig. 8a). Fourteen days after fertilization, the
oral surface was no longer evidently sunken, and the larval
para-arms and anterior yolky mass have moved away from
the oral surface toward the aboral surface of the juvenile
(Fig. 8b, c).
At 101 h post-fertilization, a hydropore was evident on
the dorsal surface of the larva (Fig. 7a). The location of
this pore was near the median side of the base of right
anterior para-arm. Sections of 101-h-old larvae showed
that the hydropore was joined to the hydrocoel via a canal
lined by a thick epithelium (Fig. 7e). In sections of younger
larvae (88.5 h), this hydroporic canal invaginated from
the larval surface but was not yet joined to the coelomic
cavities. Sections of 14-day larvae showed the hydropore
connected to the hydrocoel by a stone canal (Fig. 8c, e).
Also by this stage, epineural folds had joined to form an
epineural sinus (Fig. 8d, e).
Observations on the calcitic skeleton
Larval stages at 58, 63, 75.5, 88.5. and 101 h after fer-
tilization were cleared to look for calcareous skeletal spic-
ules within developing embryos. No evidence of calcifi-
cation was seen in 58- and 63-h specimens, even though
the latter had begun to form the para-arms (Figs. 4f and
9a). The first evidence of calcification was found in 75.5-
h specimens (Fig. 9b). In these, para-arms were well
formed, and podial bulges had just begun to form. One
calcareous plate-like ossicle was embedded in the base of
each para-arm. In the more advanced 75.5-h specimen of
the two observed, a fifth calcareous ossicle was present
and located centrally between the four para-arms (Fig.
9b). In 88.5-h specimens, the five ossicles had grown into
plates, and those in the para-arms had formed fenestrated
rods that projected toward the distal ends of the para-
ECHINOTHURIOID DEVELOPMENT, REVISITED
23
Figure 7. A 101-h larvae of Asthenosoma ijimai. a. Light micrograph of dorsal side of live specimen.
Note the hydropore (Hp) and four para-arms (to right). The anterior end is to the left of figure. Scale bar,
0.5 mm. b. Light micrograph of ventral side of live specimen. The anterior end is to the right of figure.
Same scale as a. c. Medial frontal section through larva shows developing internal structures and juvenile
oral region. P, podia; Rs, right somatocoel; Ls. left somatocoel; G. remnant of archenteron and future gut.
Same scale as a. d. Close-up of juvenile oral region, with podia (P). epineural folds (Ef), radial canals of
water vascular system (R). and left somatocoel (Ls). Scale bar. 200 pm. e. Section at the level of the hydropore
shows mvaginated canal (He). In an adjacent section, the canal joins the hydrocoel. Scale bar, 200 nm. f.
SEM of oral region of larva, shows five podia, and bulges for spines (Sp) sunken into the left larval surface.
Scale bar. 0.5 mm. g. Close-up SEM of oral region showing inward movement of epineural folds (Ef) between
podia (P). The infolding epidermis is strongly ciliated whereas the original floor of the oral region is sparsely
ciliated. Scale bar. 200 urn. h. High magnification view of a single epineural fold (Ef) moving between two
adjacent podia (P). Scale bar. 50 ftm.
arms (Fig. 9c). These rods were particularly well developed
in the two right para-arms and had just begun to form in
the two left para-arms. The centrally located plate showed
no evidence of an attached rod. Each of the calcined skel-
etal plates, with or without rods attached, behaved opti-
cally like a single crystal when rotated through polarized
light (Fig. 9d-f). This observation confirmed the structural
appearance that plates with attached rods were a single
skeletal unit. Also in the 88.5-h larvae, several other cal-
cification centers had formed and ossicles were growing
(Fig. 9c).
Calcification in 101-h larvae was even more developed
(Fig. 9g). These larvae had well-developed podial buds
(Fig. 7b) and, on one specimen, the buds for spines were
developing on the circumference of the juvenile oral sur-
face (see Fig. 7f). As with earlier stages, fenestrated rods
24
S. AMEMIYA AND R. B. EMLET
Figure 8. Later stages of larval development of Asthenoxoma ijimai. For all specimens, the anterior end
is to the right of figure, a. Ventral side of live specimen nine days after fertilization. Scale bar, 0.5 mm. b.
Ventral side of live specimen 14 days after fertilization. The larval para-arms and anterior yolky mass have
been contorted toward the juvenile aboral surface. P, podia; Sp, spines. Same scale as a. c. Fourteen-day
post fertilization, approximate frontal section at the level of the hydropore and hydroporic canal (He). Same
scale as a. d. Close-up of juvenile oral region showing epineural sinuses (Es), gut (G), water vascular system
(W), radial canal (R), and podia (P). Compare with e. Scale bar. 200 ^m. e. SEM of partially sectioned
specimen showing similar structures as seen in d. He, hydroporic canal; Rs, right somatocoel (aboral part
of body cavity); Ls, left somatocoel (oral part of body cavity). Scale bar, 200 pm.
were associated with plates in the para-arms and not with
other ossicles. In one larva, each of the spine buds con-
tained a growing spicule. In this same larva, the two os-
sicles of the left para-arms and three additional ossicles
formed a circle beneath the juvenile oral surface that rep-
resented the five ocular plates of the adult skeleton.
Discussion
Larval structure o/'Asthenosoma ijimai
Our re-examination of the larval development of As-
thenosoma ijimai has demonstrated several morphological
features that were not reported in the initial study of this
species. Amemiya and Tsuchiya (1979) reported that the
early post-gastrula of .1. ijimai resembled an early bi-
pinnaria and not a prism larva. That study also reported
the appearance of para-arms later in development and
distinguished these projections from pluteus larval arms,
because the former apparently lacked larval spicules and
apparently arose from different regions of the larva. On
this basis Amemiya and Tsuchiya concluded that, during
development, A. ijimai passes from the gastrula stage to
metamorphosis without showing any evidence of a pluteus
larval form. They also concluded that the development
of A. ijimai represents a second example of direct devel-
opment (sensu Hyman, 1955) for an echinoid, the first
being that of Heliocidaris erythrogramma (development
originally described by Mortensen, 1921, but also by Wil-
liams and Anderson, 1975). Amemiya and Tsuchiya
(1979) identified the surface on which para-arms arose in
embryos of Asthenosoma as the ventral surface because
of its resemblance to the ventral (oral) surface of early
bipinnaria larvae of asteroids. Amemiya and Tsuchiya
( 1979) also incorrectly stated that the five primary podia
were on the ventral surface, although Amemiya (1980)
reported that primary podia arise lateral to the ventral
surface. In the present study, the surface on which the
para-arms arose has been identified as the dorsal surface
based on observations of internal structures and on com-
parison with the primitive pluteus morphology. In this
new orientation, the primary podia form on the left side
of the larva.
A number of newly observed structures and their po-
sitions lead us to reinterpret the larval development of
ECHINOTHUR1OID DEVELOPMENT. REVISITED
25
Figure 9. Skeletal development in various, cleared stages of larvae of Asthenosoma ijimai. All larvae
are viewed from the dorsal side in partially polarized light, a. A 63-h larva shows no evidence of calcareous
skeletal elements. Scale bar, 0.5 mm. b. Two 75.5-h specimens show the very first signs of skeletal development.
One calcareous element is associated with each para-arm. The specimen on the right was an additional
calcareous element centrally located between the para-arms. Same scale as a. c. A 88.5-h larva with continued
skeletal development. Each calcareous element associated with a para-arm has formed a plate-like ossicle
and shows substantial or initial formation of a rod attached to the plate. Other calcification centers have
also begun. Same scale as a. d. Close-up of plate-like ossicle and rod from right posterior para-arm of a 88.5-
h larva. Scale bar. 100 ^m. e. Another plate-like ossicle and rod from a 101-h larva. Same scale as d. f.
Central plate-like ossicle without an associated rod from a 101-h larva. Scale bar, 100 ^m. g. A 101-h larva
with manv calcification sites. Same scale as a.
Asthenosoma ijimai as that of a highly modified pluteus
larva. The two pair of bilaterally symmetrical para-arms
arising from posterior and dorsal parts of embryo, each
one containing a calcareous, fenestrated skeletal element,
appear to be vestigial larval arms. We reject an alternative
interpretation that the fenestrated rods are juvenile spines,
because the spines form in association with plates that
are separate elements (Gordon, 1926a, b). Because the
skeletal elements are fenestrated, we interpret the para-
arms as reduced post-oral and postero-dorsal arms (the
26
S. AMEMIYA AND R. B. EMLET
1st and 3rd pairs of arms) of a pluteus. Fenestrated skeletal
rods are only known for these arm pairs in pluteus larvae
( Mortensen, 1 92 1 ; Emlet, 1 982). In typical plutei, the sec-
ond pair of arms to form is the anterolateral pair that
always contains simple calcareous rods (Mortensen, 1 92 1 ).
Each anterolateral rod is an outgrowth from the pair of
spicules that also form the postoral rods and body skel-
eton. The postoral rods are so reduced in A. ijimai that
anterolateral rods are absent.
There are also several differences in the early formation
of fenestrated spicules in a pluteus and those in A. ijimai.
( 1 ) In the pluteus, a fenestrated rod grows from a triradiate
spicule (Okazaki, 1975) and later elaborates a plate at its
proximal base (Emlet, 1985. and unpub. obs.). In contrast,
skeletal elements in A. ijimai form proximal, reticulate
plate-like ossicles that later form reduced, fenestrated rods.
(2) In a pluteus, calcareous rods extend and consequently
the arms elongate (e.g., Okazaki, 1975); in larvae of A.
ijimai, para-arms are already present before spicules
elongate. In actuality, formation of arm buds in the ab-
sence of spicules can still occur in plutei (Yasumasu et
a/., 1985; Emlet, pers. obs.) indicating that the epidermis
of the arm regions is apparently distinct prior to its as-
sociation with spicules. This last observation is consistent
with the formation of arm buds in A. ijimai.
Additional support for the identification of the para-
arms as homologues of the first and third arm pairs of a
pluteus larva comes from the following evaluation of arm
position. Rather than being directed anteriorly (in the di-
rection of swimming) as they are for a pluteus larva, arms
and their associated skeletal elements are reflected dorsally
and posteriorly at the surface of the very large yolky larva
(Fig. 10). In echinoids with plutei, the gastrula forms a
prism larva when rods of the first pair of larval spicules
lengthen into postoral, body, and anterolateral rods and
deform the ectoderm (Horstadius, 1939; Okazaki, 1975).
The prism's ventral surface (defined by the association of
the archenteron tip with that surface) flattens to become
the pluteus oral surface; the prism's dorsal surface (op-
posite the ventral surface) distends to become the aboral
surface, terminating at the posterior end of the pluteus.
During the prism stage, the ciliated band forms and serves
as a landmark dividing oral and aboral ectoderm (c.f.,
Davidson, 1 986). The postoral arms grow anteriorly from
the positions lateral to the blastopore. Late in the four-
armed stage, a second pair of triradiate spicules appears
at dorsolateral edges of aboral surface near the ciliated
band (see Fig. 10), and these form the (usually) fenestrated
posterodorsal rods (e.g., Mortensen, 1921; Okazaki, 1975).
The postoral and posterodorsal arms thus extend the cil-
iated band anteriorly and are located at the edge of the
concave oral and convex aboral ectoderm. If the posterior
end of the pluteus were not convex, and if the aboral
surface lay in one plane, the positions of the postoral and
DORSAL VIEW
PO{
PD
Figure 10. Schematic of a larva ofAs/lienosonia ijimai and a pluteus
(Strongylocentrotusfranciscanus) viewed from dorsal and left lateral ori-
entations. In A ijimai the para-arms are reflected posteriorly and contain
reduced skeletal elements. These bilaterally symmetric arms and spicules
are in positions that can be considered homologous with the postoral
(PO) and posterodorsal arms (PD) of the pluteus. The anterolateral arms
and rods (AL) have been lost in A. ijimai- Hp, hydropore; S, stomach.
posterodorsal arms of a pluteus would conform with the
para-arms of Asthenosoma ijimai (see Fig. 10). This de-
scription is consistent with the hypothesis of homology
between the identified arms and skeletal elements in plutei
and larvae of A. ijimai.
The position of another newly observed structure, the
hydropore, is also consistent with and supports this in-
terpretation of vestiges of pluteus larval development. In
both pluteus larvae and those of Asthenosoma ijimai, the
hydropore opens medially to, and anterior of, the bases
of the posterodorsal arms (Figs. 10, 7a). A clear difference
is, however, that the hydropore opens just after coelom
formation in pluteus development and it opens only after
advanced coelomic development in A. ijimai.
If the boundary between oral and aboral ectoderm has
remained associated with the epidermal regions of the
arms, this reinterpretation of the larval form of Asthe-
nosoma ijimai implies that the large, rounded, anterior
end of the larva is covered by oral ectoderm and that
aboral ectoderm may be restricted to that region associated
with the para-arms. For A. ijimai. there may be a reversal
in the relative area (and shape) of the oral ectoderm and
aboral ectoderm compared to that in plutei (Fig. 10). It
ECHINOTHURIOID DEVELOPMENT, REVISITED
27
may be possible to test this hypothesis with cell lineage
studies or with immunocytochemical probes to transcripts
of the Cylll actin gene or the Spec gene, which are specific
to aboral ectoderm in plutei of Strongylocentrotus pur-
piiratus (Cox et ai. 1986; Davidson, 1986). Enlargement
of the oral ectoderm and reduction of aboral ectoderm
has been demonstrated in cell lineage studies of Helioci-
daris erythrogramma (Wray and Raff, 1990). Further
work will be required to determine whether this apparent
similarity represents a new case of parallelism in echinoid
developmental patterns.
Comparisons between pluteus development and
modified development
Even though larvae of Asthenosoma ijimai retain sev-
eral reduced pluteus structures, several other features are
partially convergent with other echinoid species that have
modified development. Developmental comparisons
among species that form plutei, A. ijimai, and other spe-
cies with modified development allows inferences about
morphogenetic changes that may occur during evolution
from pluteus development to highly modified (e.g.. direct)
development.
An equal fourth cleavage, documented here for As-
thenosoma ijimai, is a common feature of species with
highly modified development and is correlated with the
production of a large number of mesenchyme cells (Raff,
1987; Parks et al, 1989). Raff (1987) suggested that the
large number of mesenchyme cells is a requirement for
acceleration of development of the adult rudiment. For
A. ijimai. only a fraction of the large number of blasto-
coelic cells become skeletogenic and only after a delay
relative to species with feeding larvae. A large number of
mesenchyme cells is also produced from the relatively
large micromeres of an unequal fourth cleavage by em-
bryos of Peronella japonica, and some of these cells also
produce larval skeletal rods (Okazaki and Dan, 1954;
Okazaki, 1975). These comparisons suggest that the loss
of the expression of larval skeleton is independent of, and
follows amplification of, the cell lineage that putatively
produces adult skeleton.
The growth and behavior of the archenteron and coe-
loms of Asthenosoma ijimai appears to be intermediate
between that of species with pluteus development and
that of the other species with modified development. In
species with feeding larvae, the archenteron grows into
the blastocoel, reaching approximately % of the distance
toward the animal pole prior to bending toward, and at-
taching to, the blastocoel wall where the larval mouth
forms. In most species for which modified development
has been described, the archenteron invaginates less than
halfway into the blastocoel: Peronella japonica (Morten-
sen, 1921; Okazaki, 1975); Heliocidaris erythrogramma
(Williams and Anderson, 1975; Wray and Raff, 1989);
Phyllacanthus parvispimis (Parks et al., 1989). In Asthen-
osoma ijimai, the archenteron invaginated as much as %
of the way into the blastocoel. Unlike what has been re-
ported for other species with modified development, in
A. ijimai the tip of the archenteron curved toward the
ventral surface of the blastocoel and subsequently under-
went torsion to the left side of the larva.
Enterocoely, where left and right coelomic pouches are
budded off completely from the anterior end of the arch-
enteron, is the only means of coelom formation known
in pluteus development (e.g., Okazaki, 1975). Both of
these pouches divide again to form anterior axocoelic and
posterior somatocoelic sacs. The left axocoel subsequently
grows a canal to form the dorsal hydropore, and another
extension of this sac forms the left hydrocoel (Hyman,
1955). Development of bilateral coelomic pouches fol-
lowed by posterior growth and formation of somatocoels
also occurs in Asthenosoma ijimai, yet in a distinctive
way: two outpocketings from the archenteron form the
left and right somatocoels, and precocious hydrocoelic
lobes grow from the tip of the archenteron (Figs. 4; 5c,
d). No obvious axocoelic sacs and no hydroporic canal
are formed during this sequence. Later, a hydropore does
form and joins with the water vascular system. In other
species with modified development, the coelomic pouches
are usually produced in pairs at the tip of the archenteron,
with one sac substantially larger than the other (e.g., Wil-
liams and Anderson, 1975; Wray and Raff, 1989). Several
species with highly modified development are reported to
form additional coelomic sacs by shizocoely from aggre-
gated mesenchyme cells (e.g., Williams and Anderson,
1975; Schatt, 1985). These patterns suggest that a tran-
sition from enterocoely to a combination of enterocoely
and shizocoely may take place only after considerable
modification of development has already occurred. For
species with modified development in general, detailed
descriptions of how coelomic pouches give rise to different
coelomic sacs are currently lacking, and additional studies
are needed.
The orientation of the adult oral-aboral axis relative to
the plane of bilateral symmetry of the pluteus is conserved
in several species with modified development, but is ap-
parently lost in others. Loss of symmetry is not related to
the degree of loss of pluteus features. Evidence for reten-
tion of pluteus larval symmetry and its relation to juvenile
rudiment formation has already been presented for As-
thenosoma ijimai. In Phyllacanthus imperialis, with two
pairs of larval arms, a reduced preoral region, and no
larval mouth, the juvenile oral surface forms on the left
side of the larval body (Olson et al., 1988). A reduced
bilateral symmetry is also present in Heliocidaris erythro-
gramma, which has one coelomic pouch (the left one)
larger than the other (Williams and Anderson, 1975) and
28
S. AMEMIYA AND R. B. EMLET
a bilaterally symmetric (larval) serotonergic ganglion
(Bisgrove and Raft", 1989). In H. erythrogramma, the
coincident arrangement of these two sources of bilaterality
provides evidence for conservation of the adult oral-aboral
axis (Bisgrove and Raft", 1989). Departures from conser-
vation of relative positions of larval and adult axes occur
for three other species. Peronella japonica and P. rubra
(with very similar development) have bilaterally sym-
metric larvae, but the juvenile rudiment forms centrally,
with the juvenile oral surface directed anteriorly and dor-
sally (Okazaki, 1975; Amemiya and Emlet, unpub. obs.).
Loss of the primitive larval-adult arrangement of axes may
be related to the retention of one pair of larval arms (pos-
torals) and the loss of the more dorsal pair. Retention of
only one well-developed pair of larval arms may force the
rudiment to develop on the dorsal side, whereas in the
absence or reduction of both pairs (e.g., H. erythro-
gramma, A. ijimai), or retention of both pairs (P. impe-
rialis), the juvenile oral surface would not be shifted. This
mechanistic hypothesis does not apply to Phyllacanthus
parvispimts which lacks a larval skeleton and has bilobed,
asymmetric coeloms which do not coincide with the ori-
entation of the serotonergic neurons (Park et a/., 1989).
Further morphological studies of the formation and
growth of the archenteron and coeloms of both P. japonica
and P. parvispimts are needed to determine how primitive
larval and adult oral aboral axes have been rearranged.
Eitechinoid characters in echinothurioid development
Early ingression of cells from the blastular wall in As-
thenosoma ijimai (Fig. Id) is comparable to primary mes-
enchyme ingression, which occurs prior to gastrulation in
other euechinoids and is different from the later ingression
known among the cidaroids (Schroeder, 198 1 ).
Beginning with the first appearance of podia and con-
tinuing until the adult skeleton is well-developed (Fig. 8b,
c), the juvenile oral region of Asthe nosoma ijimai sinks
into the surface of the larva (Figs. 7b, c, f; 8a). Within
this indentation, both podia and oral spines grow. Though
there is no early invagination and enclosure of the juvenile
oral surface that could be clearly identified as an amniotic
invagination, the strong indentation may be a morpho-
genetic process equivalent to vestibule formation. Other
species with modified development either have or lack an
amniotic invagination, consistent with their phylogenetic
position as cidaroids or euechinoids (see Parks et ai.
1989), and this sunken condition is, therefore, not simply
due to the yolkiness of the larva. A comparison of our
figures with those of the cidaroid Phyllacanthus pani-
spimis. which lacks an amniotic invagination, shows that
the oral surface of A. ijimai is considerably more indented
than that of the cidaroid (Parks ci al.. 1989, fig. 3f). Our
observations raise the possibility that a partial, possibly
primitive, form of an amniotic invagination may be pres-
ent in echinothurioids.
Parks et al. (1989) also examined sectioned material of
A. ijimai and concluded that an amniotic invagination
was absent. These authors hypothesized that an amniotic
invagination arose in the euechinoid lineage after the
echinothurioid branch. They hypothesized further that,
because the two most primitive lineages of echinoids, the
cidaroids and echinothurioids, lacked an amniotic invag-
ination, the absence of this character was primitive for
echinoids. Our observations suggest that their first phy-
logenetic hypothesis may not be accurate, but our findings
are consistent with the hypothesis that the amniotic in-
vagination is a derived character in euechinoids. Based
on comparisons of the fate of larval epidermis among
echinoderm classes, Emlet (1988) suggested the same hy-
pothesis that the primitive condition for echinoids is the
absence of an amniotic invagination.
The formation of epineural sinuses in Asthenosoma iji-
mai matches very closely the original descriptions of the
same process occurring within the amniotic invagination
of other euechinoids [compare Fig. 7f-h with original text-
figs, e-h of von Ubisch, 1913 (text-figs, f, g reprinted in
Hyman, 1955, p. 497)]. By contrast, epineural sinus for-
mation in this echinothurioid differs from that described
for the cidaroid, Eucidaris thouarsi (Emlet, 1988). In E.
thouarsi, epineural folds were present, but not clearly ev-
ident when observed by SEM. Sections of E. thouarsi
showed epineural folds closely adhered to the developing
juvenile oral surface on the left side of the larva, whereas
in euechinoids, the epineural folds were not so closely
adhered. Emlet (1988) hypothesized that the pattern in
E. thouarsi might reflect a different mechanism of epi-
neural fold movement (from that in euechinoids) but
might also result from the open condition of the surface
upon which this process occurs in E. thouarsi. The largely
open nature of the oral surface in A. ijimai, and the dis-
tinctly euechinoid appearance of its epineural folds, sug-
gest that the open condition of the epineural folds in E.
thouarsi was not a cause for their appearance. This ob-
servation dismisses Emlet's (1988) hypothesized expla-
nation for convergence of the epineural folds ofE. thouarsi
and the ophiuroid, Ophiopholis aculeata (Olsen, 1942)
but leaves standing the hypothesis that cidaroids and
ophiuroids have similar means of epineural sinus for-
mation (Emlet, 1988).
Conclusions
In this re-examination of the larval morphogenesis of
Asthenosoma ijimai, evidence has been presented to show
that A. ijimai has retained previously unrecognized, re-
duced pluteus characters. As such, this larval form is the
most reduced pluteus yet described, being considerably
ECHINOTHURIOID DEVELOPMENT, REVISITED
29
more modified than larvae of Phyllacanthus imperialis
(Olson et ai, 1988) and Peronella japonica. This contri-
bution brings to three the number of lineages with mod-
ified development and with pluteus characters that are
retained to varying degrees. In contrast, four lineages
(Phyllacanthus parvispinus. Heliocidaris erythrogramma,
a temnopleuroid, and Abalns species) have lost most, if
not all, primitive larval characters. (The genus Phyllacan-
thus stands alone as being represented in both groups, but
it is not known whether non-feeding development has
evolved independently for the two species: P. imperialis
and P. pun'ispinus.) Comparative experimental and cell
lineage information about species with reduced larval fea-
tures must now be collected if we are to determine 1 )
whether there is a common, possibly convergent, theme
of developmental changes, or 2) whether modifications
to cleavage and cell lineage fates are additional changes
occurring after the loss of feeding and the reduction of
the pluteus form. The detailed description presented here
adds to the growing collection of comparative data on
modified development; but it also indicates that the basic
morphological changes occurring in other species with
modified development — including two that are already
well studied, Heliocidaris erythrogramma and Peronella
japonica — ought to be re-examined.
Acknowledgments
This research was supported by grants from the Japa-
nese Society for Promotion of Science (to SA and RBE)
and the United States National Science Foundation (BSR-
9058139 to RBE). We would like to thank E. Arakawa
for technical assistance, M. McFall-Ngai for allowing us
to use her digitizing software, I. Lagomarsino for digitizing
so many serial sections, and R. A. Raff for providing the
B2C2 antibody. We are also grateful to S. Smiley for advice
on clearing opaque larvae. Comments of V. Morris and
two anonymous reviewers helped improve the manuscript.
Literature Cited
Amemiya, S. 1980. Metamorphosis of an echinothurioid sea urchin.
Aslhenosoma ijimai. Pp. 74-78 in An Outline of Modern Biology
Vol. I IB. I. Yasumasu et al. eds. Kakayama-shoten, Tokyo, (in Jap-
anese)
Amemiya, S., and T. Tsuchiya. 1979. Development of the echinothurid
sea urchin Aslhenosoma iiunai Mar Bioi 52: 93-96.
Armstrong, P. B., and D. Parent!. 1973. Scanning electron microscopy
of the chick embryo. De\: Bioi 33: 457-462.
Bisgrove, B. W., and R. A. Raff. 1989. Evolutionary conservation of
the larval serotonergic nervous system in a direct developing sea ur-
chin. Dev. Growth Differ. 31: 363-370.
Cox, K. H., L. M. Angerer, J. J. Lee, E. H. Davidson, and R. C. Angerer.
1986. Cell lineage-specific programs of expression of multiple actin
genes during sea urchin embryogenesis. / Mol. Bioi 188: 159-172.
Davidson, E. H. 1986. (jene Activity in Early Development. 3rd edition.
Academic Press. Orlando. FL.
Emlet, R. B. 1982. Echinoderm calcite: a mechanical analysis from
larval spicules. Bioi Bull. 163: 264-275.
Emlet, R. B. 1985. Crystal axes in recent and fossil adult echinoids
indicate trophic mode in larval development. Science 230: 937-940.
Emlet, R. B. 1988. Larval form and metamorphosis of a "primitive"
sea urchin, Eucidaris thonarsi (Echinodermata: Echinoidea: Cida-
roida). with implications for developmental and phylogenetic studies.
Bioi. Bull 174:4-19.
Emlet, R. B. 1990. World patterns of developmental mode in echinoid
echinoderms. Pp. 329-335 in Advances in Invertebrate Reproduction,
I'ol. 5. M. Hoshi and O. Yamashita. eds. Elsevier Science Publ..
Amsterdam.
Emlet, R. B., L. R. McEdward, and R. R. Strathmann. 1987. Echino-
derm larval ecology viewed from the egg. Pp. 55-136 in Echinoderm
Studies. Vol. 2, M. Jangoux and J. M. Lawrence, eds. Balkema Press.
Rotterdam.
Fell, F. J. 1976. The Cidaroida (Echinodermata: Echinoidea) of Ant-
arctica and the southern oceans. Unpublished Ph.D. dissertation.
University of Maine, Orono, Maine. 293 pp.
Gordon, I. 1926a. The development of the calcareous test of Echinus
milians. Philos. Trans. R Six.: B 214: 259-312.
Gordon, I. 1926b. The development of the calcareous test of Ecliino-
cardium cordatum Philos. Trans. R. Sac. B 215: 255-313.
Harris, P., and G. Shavt. 1984. Intermediate filaments, microtubules
and microfilaments in epidermis of sea urchin tubefoot. Cell Tissue
Res. 236: 27-33.
Henry, J. J., and R. A. Raff. 1990. Evolutionary change in the process
of dorsoventral axis determination in the direct developing sea urchin.
Heliocidaris erythrogramma. Dev Bioi 141: 55-69.
Horstadius, S. 1939. The mechanics of sea urchin development, studied
by operative methods. Bioi. Rev Camb. Philos. Soc. 14: 132-179.
1 1\ m:in. L. H. 1955. The Invertebrates: Echinodermata. I'ol. II'
McGraw-Hill, New York.
Larrain, A. 1973. Los pedicelarios globiferos y el desarrollo de Abatus
agassi:i. Bol. Soc. Bioi Concepcion 46: 53-63.
Mortensen, T. 1921. Studies of the Development and Larval Forms of
Echinoderms. G.E.C. Gad. Copenhagen.
Mortensen, T. 1936. Echinoidea and Ophiuroidea. Discover}' Rep 12:
199-348.
Okazaki, K. 1975. Normal development to metamorphosis. Pp. 177-
232 in The Sea Urchin Embryo Biochemistry and Morphogenesis,
G. Czihak, ed. Springer-Verlag. Berlin.
Okazaki, K.. and K. Dan. 1954. The metamorphosis of partial larvae
of Peronella japonica Mortensen. a sand dollar. Bioi Bull. 106: 83-
99.
Olsen, H. 1942. Development of a brittlestar Ophiopholis aculeata.
with a short report on the outer hyaline layer. Bergens Museum Aar-
bok. .\atur. 6: 1-107.
Olson, R. R., J. L. Cameron, and C. M. Young. 1988. Larval devel-
opment of the pencil urchin Phyllacanthus imperialis: a lecithotrophic
pluteus. P. 807 in Echinoderm Biology. Proc 6th Int'l Echinoderm
Conl. Victoria. R. D. Burke. P. V. Mladenov, P. Lambert. R. L.
Parsley, eds. Balkema. Rotterdam.
Parks, A. L., B. A. Parr, J.-E. Chin, D. S. Leaf, and R. A. Raff. 1988.
Molecular analysis of heterochronic changes in the evolution of direct
developing sea urchins. / Evol. Bioi. 1: 27-44.
Parks, A. L., B. \V. Bisgrove, G. A. Wray, and R. A. Raff. 1989. Direct
development in the sea urchin Phyllacanthus parvispinus (Cidaroidea):
phylogenetic history and functional modification. Bioi. Bull 177:
96-109.
Raff, R. A. 1987. Constraint, flexibility, and phylogenetic history' in
the evolution of direct development in sea urchins. Dev Bioi 119:
6-19.
30
S. AMEM1YA AND R. B. EMLET
Raff, R. A., B. Parr, A. Parks, and G. Wray. 1990. Radical evolutionary
change in early development. Pp. 71-98 in Evolutionary Innovations,
M. H. Nitecki, ed. University of Chicago Press, Chicago, IL.
Richardson, K. C, L. Jarrett, and E. H. Finke. 1960. Embedding in
epoxy resin for ultrathin sectioning in electron microscopy. Slain
Technol. 35:313-323.
Schatt, Ph. 1985. Developpement et croissance embryonnaire de
1'oursin incubant Abatus cordatus (Echinoidea: Spatangoida). These
de Doctoral de 1'Universite Pierre et Marie Curie. 151 pp.
Schatt, Ph. 1988. Embryonic growth of the brooding sea urchin Ahatux
cordatus. Pp. 225-228 in Echinodcnn Biology. Proc. Mh Inl '/ Echi-
noderm Con/.. Victoria, R. D. Burke, P. V. Mladenov, P. Lambert,
and R. L. Parsley, eds. Balkema. Rotterdam.
Schroeder, T. E. 1981. Development of a "primitive" sea urchin (Eu-
cidaris tribuloides): irregularities in the hyaline layer, micromeres,
and primary mesenchyme. Biol liuli 161: 141-151.
Smith, A. B. 1984. Echinoid Paleobiology. George Allen and Unwin,
London.
Strathmann, R. R. 1978. The evolution and loss feeding larval stages
of marine invertebrates. Evolution 32: 894-906.
Ubisch, L. von. 1913. Die entwicklung von Slrongylocentrotus livuhis.
(Echinus microtuberculatus, Arbacia pustulosa). Z. Wiss. Zoo/. 106:
409-448.
\\ illiams, D. H. C., and D. T. Anderson. 1975. The reproductive system,
embryonic development, larval development and metamorphosis of
the sea urchin Heliucidaris erythragramma (Val.) (Echinoidea: Echi-
nometndae). Aust. J Zoo/. 23: 371-403.
Wray, G. A., and R. A. Raff. 1989. Evolutionary modification of cell
lineage in the direct-developing sea urchin He/iociiiaris erythro-
gramma. Dev Biol. 132: 458-470.
Wray, G. A., and R. A. Raff. 1990. Novel origins of lineage founder
cells in the direct-developing sea urchin Heliocidaris erythrogramma.
Dev. Biol. 141:41-54.
Yasumasu, I., K. Mitsunaga, and V. Fujino. 1985. Mechanism forelec-
trosilent Ca:+ transport to cause calcification of spicules in sea urchin
embryos. Exp. Cell Res 159: 80-90.
Reference: Biol. Bull 182: 31-40. (February, 1992)
Purification and Biochemical Characterization
of the Nuclear Sperm-Specific Proteins of
the Bivalve Mollusks Agriodesma saxicola1
and Mytilimeria nuttalli
JUAN AUSIO
Department of Biochemistry and Microbiology, University of Victoria,
I 'ictoria. British Columbia I '8 W 3P6. Canada
Abstract. The proteins from the nuclei of the sperm
from two different species of the subclass Anomalodes-
mata of the class Bivalvia have been analyzed for the first
time. In both instances — Agriodesma saxicola (Baird,
1863) and Mytilimeria nuttalli (Conrad, 1837)— the
compositional pattern is very similar. The sperm chro-
matin is organized by a major protamine-like PL-I protein.
As in all PL-I, this protein has a trypsin-resistant core. In
both species analyzed, PL-I contains cysteine residues that
account for the presence of the monomer (M) and dimer
(D) forms observed in the total nuclear HC1 extracts. The
molecular mass of these proteins is 21,000 Da in A. sax-
icola, and 25,000 Da in M. nuttalli. All of the specimens
of A. saxicola analyzed were hermaphrodites. As a result,
the nuclear sperm-specific proteins from several preparations
were readily and extensively degraded by protease activity
from the oocytes. Such degradation was always observed
when cross contamination between the two gonadal tissues
accidentally occurred during protein extraction.
Introduction
The presence of a highly specialized histone HI (PL-I
protein) seems to be a common feature of the nuclear protein
composition of the sperm of bivalve mollusks (Jutglar et al..
1991). Despite the structural heterogeneity of the sperm
proteins within this taxonomic group (Ausio, 1 986; Zalensky
and Zalenskaya, 1980), a PL-I protein has been identified
in each of the species analyzed in detail so far. Like histone
Received 18 June 1 99 1 ; accepted 25 November 1991.
' This species is more commonly referred to as Entodesma saxicola,
but see Bernard. 1983.
HI, this protein is soluble in diluted perchloric acid, has a
globular trypsin-resistant core, is lysine rich, and yet is com-
positionally related to the protamines (PL = protamine-like)
(Subirana et al., 1973). Of all the different nuclear sperm-
specific proteins found within a given species, PL-I is the
one with the lowest electrophoretic mobility in urea-acetic
acid gels (Ausio, 1986). The presence of a protamine-like
histone HI -like protein in bivalve mollusks may have im-
portant evolutionary implications, not only within the phy-
lum Mollusca (Subirana and Colom, 1987), but also within
other taxonomic groups.
As pointed out by Kasinsky (1989), however, only some
of the subclasses within the class Bivalvia have been an-
alyzed so far. and only a few species have been thoroughly
characterized. Nevertheless, at least one sperm-specific
histone HI (PL-I) protein has been identified: in Mytilus
californianus (by Jutglar et al.. 1991), Crassostrea gigas
(by Sellos, 1985), and Glycymeris yesonensis (by Odin-
tsova et al, 1989) (subclass Pteriomorphia); in Ensis minor
(by Giancotti et al., 1983), Spisula solidissima (by Ausio
etal., 1987) and Macoma nasuta (Ausio, 1988) (Subclass
Heterodonta); and in Anodonta piscinallis (by Rozov et
al., 1984) (Subclass Palaeoheterodonta). In some of these
species, two histone HI -like proteins have been described.
Although the subclass Heterodonta has been widely
studied (Ausio, 1986), three subclasses, according to
Barnes's (1980) classification of the bivalve mollusks, have
never been characterized: Cryptodonta, Palaeotaxodonta,
and Anomalodesmata. In the present work, we have an-
alyzed and characterized the sperm-specific proteins of
two species within the subclass Anomalodesmata and have
shown that each contains a highly specialized histone H 1 -
31
32
J. AUSIO
like (PL-I) protein that is the major protein component
of the nuclei of the sperm of that organism.
Materials and Methods
Living specimens
Specimens of Agriodesrna saxicola and Mytilimeria
nuttalli were collected along the west coast of Vancouver
Island. British Columbia, Canada, by SCUBA divers from
the Biology Department at the University of Victoria.
Nuclei preparation and protein extraction
Isolation of sperm nuclei and HC1 crude extraction of
the nuclear basic proteins was performed as described
elsewhere (Ausio, 1986). Briefly, after carefully opening
the shell, a small incision was made in the gonadal tissue,
and the spontaneously released sperm were resuspended
in NaCl 0.15 M, Tris-HCl 20 mM pH 7.6, 0.2 mM PMSF
(Phenylmethylsulphonyl Fluoride) (buffer A). Because,-!.
saxicola is hermaphroditic, its sperm would sometimes
be contaminated with oocytes released accidentally from
the intimately associated ovaries (see below for discussion).
The sperm suspension was centrifuged at 3000 X g for
10 min in a SS-34 Sorvall rotor at 4°C. The pellet obtained
was homogenized in buffer A containing 0.5% TritonX-
100. After standing for 10 min on ice, the suspension was
spun down under the same conditions as before. This step
is meant to solubilize most of the cytoplasmic membranes,
including the sperm flagella and the acrosome. Notice
that this step will also expose the sperm nuclei to egg
lysates in those samples of A. saxicola contaminated with
oocytes; such cytoplasmic contamination may be respon-
sible for the protein degradation observed under these
circumstances. The detergent-treated pellet was imme-
diately homogenized in 0.4 N HC1. Solubilization was
continued for 2 h under stirring at 4°C. Finally, the sus-
pension was centrifuged at 12,000 X g for 10 min at 4°C,
and the acid extract was precipitated with 6 volumes of
acetone, overnight, at -20°C.
Gel electrophoresis
Polyacrylamide gel electrophoresis was carried out on
urea-acetic acid gels, as described elsewhere (Ausio, 1986).
Protein purification and fractionation
Ionic exchange chromatography was carried out on a
10 X 100 mm Protein-Pak SP 8HR column from Waters-
Millipore as described elsewhere (Mogensen el a/.. 199 1 ).
Gel nitration was carried out on a 10 X 300 mm FPLC
Superose 12HR 10/30 column from Pharmacia. The elu-
tion buffer was 6 M guanidinium chloride (Gdn-HCl;
Schwarz/Mann Biotech), 50 mM Tris-HCl pH 7.6.
Reverse phase high pressure liquid chromatographv
(HPCL)
HPLC was carried out on a 5 n (25 X 0.46 cm) Vydac
C4 column, with 0. 1% trifluoroacetic acid (TFA) as eluant
with different acetonitrile gradients.
Determination of the molecular mass
The molecular mass of each protein was determined
by gel nitration under denaturing conditions on a Superose
12HR column in the presence of 6 M Gdn-HCl (see
above). Several protamines, protamine-like proteins, and
histones of known molecular mass were used as standards:
PL-I from Spisula solidissima (Mr: 33,500 Da) (Ausio
and Subirana, 1982a); Histone HI from calf thymus (Mr:
22,000 Da) (DeLange, 1976); PL-Ill (01) from Mytilus
edulis (Mr: 9600 Da) (Ausio and Subirana, 1982b); and
unfractionated salmine from Oncorhynchus sp. (Mr; 4300
Da) (Ando et al, 1973). Histone HI from calf thymus
was purchased from Worthington, and salmine (sulfate
form) was obtained from Sigma. The rest of the prot-
amines were prepared in my laboratory. Globular proteins
were also used as a molecular mass markers: bovine serum
albumin (Mr: 68,000); ovoalbumin (Mr: 46,000); chy-
motrypsinogen A (Mr: 25,000); and ribonuclease A (Mr:
13,200). These proteins were purchased from Pharmacia;
all of them were subjected to performic acid oxidation
before being applied to the column (see below). Vitamin
B12 (Mr: 1350) was purchased from Sigma. For the es-
timation of the molecular mass, the column was calibrated
with the above standard proteins, and a plot of Kav versus
log Mr was constructed (Mr = molecular mass; Kav
= distribution coefficient).
v, - v0
where vo and V, are the void and total volume of the column,
and Ve = the elution volume of a given protein. Blue
Dextran and dansyl-L-alanine (Sigma) were used to de-
termine V0 and V, experimentally. The proteins of un-
known Mr were mixed with the protein standards and
run together through the column. Their molecular masses
were estimated by interpolation of their Kav values on the
best fitting line of the calibration plot.
Amino acid analysis
Amino acid analysis was carried out on an Applied
Biosystems model 420A derivatizer-analyzer system. The
hydrolysis was carried out in gas-phase 6 N HC1 and 1%
phenol under an argon atmosphere at 165°C, for 1 h, 2
NUCLEAR PROTEINS FROM THE SPERM OF BIVALVE MOLLUSKS
33
h, and 4 h, the final amino acid composition was obtained
by extrapolation of the data to zero time. So that cysteine
could be quantified, all protein samples were pyridyl-
ethylated before hydrolysis, as described below.
Chemical modification of proteins
Reduction of SH groups. The SH groups of cysteine
were reduced as described by Kuehl (1979). Briefly, the
proteins, at 1 mg/ml in 6 M urea 20 mAf Tris-HCl pH
7.6, were reduced in the presence of 8% |8-mercaptoeth-
anol for 3 h at room temperature.
Oxidation ofSH groups. Oxidation was carried out un-
der the same buffer conditions as above, but in the pres-
ence of 0.72 mM O-phenanthroline and 0.36 mA/CuSO4.
Performic acid oxydation. Performic acid was prepared
according to Hirs (1967). For the oxidation, 1-mg aliquots
of protein were dissolved in 0.5 ml of performic acid,
which had been previously cooled on ice. The reaction
was allowed to proceed for 4 h in an ice bath in capped
tubes. The sample was then resuspended in a 25-fold ex-
cess of HPLC grade distilled water and lyophilized.
Cysteine pyridylethylation. Proteins were pyridyleth-
ylated, providing for a quantitative estimate of cysteine
in the amino acid analysis. The procedure used was as fol-
lows: proteins (=1 nanomol) were dissolved in 44 ^1 of
6.8 Murea, 60 mM Tris-HCl, 1.25 mA/EDTA (pH 7.6),
and 2.3% /3-mercaptoethanol. The solution was incubated
for 3 h at room temperature in the dark. Subsequently, 8
CE AS
HIST
0 I II
X T J-
IV SA
J U
PL- 1
PHI
PL- 1
PL-IV
PR
Figure 1 . Urea acetic acid PAGE analysis of the nuclear sperm-specific proteins of Agriodesma saxicola
(AS) and Mytilimeria nuttalli (MN) in comparison to a histone standard from chicken erythrocytes (CE)
and to a protamine from salmon, salmine (SA). The nuclear sperm-specific proteins of one representative
of each of the five groups (O, I, II, III. and IV) of the classification of the bivalve mollusks (Ausio. 1986) are
also shown. The representative species chosen for each group were: O: Pecten maximus; I: Spisula solidissima;
II: Ensis ensis; III: Mamma nasula. and IV: Mytilus edulis. The regions corresponding to the different
protamine-like (PL-I, PL-II, PL-Ill, and PL-IV) proteins defined in Ausio (1986) are also shown. HIST:
histone region. PR: protamine. D: dimer form. M: monomer form of the major sperm protein component
in each species. X2, Y2: possible dimer forms of the minor sperm protein components X, Y.
34
J. AUSIO
A.
0.8
0.6
0.2
0.0
ss a b c d
8
16
32 40
TIME , min
64
72
80
2.0
1.5 -
1.0 -
O
CO
0.5 2
0.0
B.
1.5
1.0
o
fO
CM
0.5
0.0
SS
L/L/UUUUUUU
10
20
30 40
TIME , min
50
60
60
50
UJ
405!
300
r-
LU
20 <
10
0
Figure 2. Fractionation of a crude 0.4 A' HCl extract from the nuclei of the sperm of Agnodesma
saxicola. (A) Ionic exchange chromatography on a (10 X 100 mm) Protein-Pak SP 8HR column. Proteins
were eluted with a linear (0-2 A/) NaCl gradient in 50 m.1/ Na-phosphate buffer (pH 6.8) at a flow rate of
1 ml/min. The inset shows the urea-acetic acid PAGE analysis of the fractions indicated. (B) Reverse-phase
HPLC on a Vydac C4 column. Elution was carried with an acetonitrile gradient in 0. 1% trifluoroacetic acid
at a flow rate of 1 ml/min. The inset shows the electrophoretic analysis of the fractionation. The lanes shown
in the inset, and the chromatogram has been aligned to match the fraction analyzed with its corresponding
position in the chromatogram. SS: starting sample.
NUCLEAR PROTEINS FROM THE SPERM OF BIVALVE MOLLUSK.S
35
A.
1.2
0.8
0.4
0.2
0.0
IL
B.
mn a b c d e f 9
_
I
?v5SHHBI^HBM^H
as a b c d e f 9
_
16
24 32
TIME ,min.
48
Figure 3. (A) Gel filtration FPLC on a Superose 12 HR 10/30 column. The elution buffer was 6 M
Gdn-HCI in 50 mA/ Tris-HCl pH 7.6. The flow rate was 0.4 ml/min. The elution profiles of HC1 nuclear
extracts from the sperm of Mytilimeria nuttalli ( ) and Agriodesma saxicola ( ) are shown together
with the elution profile (••••) of some of the standards used to calibrate this column: I: PL-I from Spisula
solidissima; II: Histone HI from calf thymus; III: PL-HI from Mytilus edulis; IV: protamine salmine; V:
vitamin B 12; VI: dansyl-L-alanine. (B) Electrophoretic analysis on urea-acetic acid gels of the fractions a,
b. c. d. e, f, g of the elution profiles of At. nuttalli and A. saxicola. mn: starting sample of A/ nuttalli. as:
starting sample of A. saxicola.
H\ of 4-vinylpyridine was added, and the reaction was al-
lowed to proceed for 2 h at room temperature. The sample
was then immediately desalted in an HPLC reverse phase
C8Vydac column, which was eluted for 5 min with 0.1%
TFA (trifluoroacetic acid), and for 20 min with a 0-70%
acetonitrile gradient in 0.1%- TFA. /3-lactoglobulin from
Applied Biosystems Inc. was used as a standard for this
procedure.
Trypsin digestion. Trypsin digestion of proteins in high
salt— 2 M NaCl, 50 mA/ Na-phosphate buffer (pH 6.8)—
was carried out as described elsewhere (Ausio el ai, \ 987).
Results
Chromatographic analysis and purification of the
sperm-specific nuclear proteins from A. saxicola and M.
nuttalli
Figure 1 shows the 0.4 N HC1 protein extracts from the
nuclei of the sperm of A. saxicola and Al. nuttalli. They
are shown in comparison to the five groups previously
established for the classification of the nuclear sperm-spe-
cific proteins of the bivalve mollusks (Ausio, 1986). In
each of the two species analyzed, two major protein bands
run in the region of the PL-I proteins (Ausio, 1986). In
addition to these proteins, 10-20% of minor protein frac-
tions X and Y, which run in the histone region, are also
observed (see Fig. 1-AS and 1-MN). This protein com-
position was sufficiently novel that the two organisms
could not, at first, be assigned to any of the protein groups
previously established in my classification of the bivalves
(Ausio, 1986). That was not surprising because they belong
to a subclass (Anomalodesmata) that had not been ana-
lyzed before. I therefore decided to purify and characterize
each of the major protein components of these organisms.
The first attempt at fractionation by ionic exchange
FPLC under non-denaturing conditions is shown in Figure
2 A. Most of the protein components coeluted in a single
multiphasic peak at around 2 M NaCl, but some protein
separation was clearly achieved as is shown in the inset
of the same figure.
36
J. AUSIO
1.0
0.8
0.6
0.4
0.2
0.0
103
10"
Mr
105
Figure 4. Calibration plot used to determine the molecular mass of
the sperm proteins determined on a superose 12HR 10/30 column under
the elution conditions described in Figure 3A. Globular O and nonglob-
ular® proteins were used as standards. 1: Vitamin B 12 (Mr: 1350 Da);
2: ribonuclease A (Mr: 13.200 Da); 3: chymotrypsingen A (Mr: 25,000
Da): 4: ovoalbumin (Mr: 46,000 Da): 5: bovine serum albumin (Mr:
68,000 Da): 1*: protamine salmine (Mr: 4,300 Da); 2*: protein PL-I
from Myl ilus editlis (Mr: 9,600 Da); 3*: histone HI from calf thymus
(Mr: 22,000 Da); 4*: Protein PL-I from Spisula solidissima (Mr: 33,000
Da). M = monomer form and D = dimer form of the major sperm
protein components of: Mytillineria mtnalli •, and Agriodesma
saxicola O.
Characterization of the sperm-specific nuclear proteins
from A. saxicola and M. nuttalli
The fractionation problems described in the preceding
section began to be elucidated when the molecular mass
of these proteins was analyzed. Figure 4 shows the cali-
bration plot used to estimate the molecular mass of pro-
teins from gel filtration analysis. The molecular mass of
the two major protein components of the sperm nuclei
in A. saxicola were: 21,000 Da for the fastest protein
component and 43,000 Da for the slowest moving frac-
tion. The values were 25,000 Da and 49,000 Da for M.
nuttalli. These results suggested a monomer-dimer rela-
tionship between the slow and the fast moving protein
fractions present in each species. To analyze this rela-
tionship, and to examine the nature of the association
phenomenon, 1 incubated the crude HC1 extracts in the
presence of either /3-mercaptoethanol or copper phen-
anthroline. Figure 5 shows the results of these treatments
in the case of A. saxicola, and identical results were ob-
tained with M. nuttalli (results not shown). The slower
moving band completely disappears under reducing con-
ditions for cysteine. Under oxidizing conditions (in the
presence of copper phenanthroline) the relative intensity
of the faster moving band (see Fig. 5b) slightly decreases,
and higher association complexes are formed (see arrow
in Fig. 5b). We are therefore dealing with the association
of the faster moving protein components. Although the
association seems to involve primarily the formation of
dimers, the number of cysteines present in the monomer
To increase the resolution in the separation, the 0.4 N
HC1 protein extracts were fractionated by reverse-phase
HPLC; the elution profile is shown in Figure 2B. Although
two peaks corresponding to each of the two major com-
ponents could be clearly separated, both of them exhibited
different amounts of what appeared to be overlapping
cross-contamination.
Size fractionation of the starting HC1 extracts by gel
filtration under denaturing conditions in the presence of
6 A/guanidinium chloride (Gdn-HCl) is shown in Figure
3. Although the peaks could not be completely resolved,
the sample was partially fractionated as shown in Figure
3B. Indeed, when some of the eluting fractions from the
different regions corresponding to the two major protein
components were pooled together and rerun under the
same conditions, two distinct peaks could then be clearly
resolved. This was used as a basis for estimating the
molecular mass of each of these protein components.
Nevertheless, when the fraction under each separate
peak was analyzed by urea acetic acid PAGE, the same
cross-contamination observed in Figure 2B was
again observed, although to a lesser extent (results not
shown).
PL-I
SS A B
. • . U
^^M
d •— •
x2
m — »»
Figure 5. Urea-acetic acid polyacrylamide gel electrophoresis of the
nuclear sperm specific proteins from Agriodesma saxicola under A: re-
ducing(6% /3-mercaptoethanol) or B: oxidizing(copper-phenanthroline)
conditions. SS = starting sample, m, d = monomer and dimer of major
sperm protein component (PL-I). X, X2 = monomer and dimer of the
minor sperm protein component. The arrowhead indicates the presence
of higher association oligomers obtained upon oxidative treatment.
NUCLEAR PROTEINS FROM THE SPERM OF BIVALVE MOLLUSKS
37
cannot be clearly ascertained from the above experiments.
Thus, although the strong tendency toward dimer for-
mation would suggest the presence only of one cysteine
per molecule, the lack of complete dimerization observed
in Figure 5b, and the presence of association complexes
higher than dimers. would strongly suggest the presence
of more than one cysteine. The presence of two cysteine
residues per molecule, which could easily form an internal
disulfide bond, would explain the incomplete polymer-
ization of the monomer, otherwise expected under the
oxidizing conditions used here (Fig. 5b).
To determine the number of cysteines, as well as to
establish the amino acid composition of the major nuclear
protein component of the sperm of A. saxicola and M.
mittalli, protein fractions such as those shown in the insets
of Figure 2A and B were pyridylethylated before amino
acid analysis. A /5-lactoglobulin sample was simulta-
neously treated and analyzed to check for the completion
of the reaction. The amino acid analyses clearly show (see
Table I) that the proteins of both A. saxicola and M. nut-
talli contain two cysteine residues per molecule. Com-
parison with the amino acid analyses of other PL proteins,
reveals the PL-I nature of the major nuclear protein com-
ponent of the sperm of the two species analyzed. Like
other PL-I proteins (Ausio, 1986; Ausio, 1988; Jutglar et
al., 1991), these have an internal trypsin resistant core
(Fig. 6).
Besides the major protein components M and D, we
have also characterized the minor component X of A.
saxicola. This protein exhibits an amino acid composition
that is almost identical to PL-I (see Table I). Although
Table I
Amino acid analysis (mol %) of the nuclear sperm-specific PL-I
proteins o/'Agriodesma saxicola PL-I (AS) and Mytilimeria nuttalli
PL-I (MN) in comparison to those of Spisula solidissima PL-I (SS)
(Ausio and Subirana. I982a) and Macoma nasuta PL-I (MC)
(Ausio. 1988)
Pl-I(AS) PL-I(MN) PL-I(SS) PL-I(MC) X (AS)
Lys
18.7
16.3
24.8
21.8
15.1
His
—
0.4
—
2.3
0.4
Arg
34.8
33.8
23.1
26.9
34.7
Asx
1.5
1.8
0.6
0.8
2.5
Thr
2.0
1.0
4.3
4.0
1.6
Ser
20.8
26.5
21.7
20.2
21.4
Glx
1.1
0.8
0.6
0.8
4.6
Pro
0.5
0.8
2.4
1.8
1.7
Gly
6.0
5.5
3.0
2 2
5.0
Ala
3.3
4.2
14.2
11.3
4.4
1/2 Cys
0.9
0.6
—
0.7
tr.*
Val
3.2
2.4
2.3
2.4
2.8
Mel
—
0.2
0.4
0.2
—
lie
1.2
0.9
0.5
1.4
1.0
Leu
3.6
3.3
1.7
2.1
2.6
Tyr
1.5
1.0
0.3
0.5
1.0
Phe
0.7
0.7
0.3
0.7
1.0
Trp
—
—
0.3
—
AS'OM OD
2o
3o 4M 4o SM 5o
* Determination carried out in the absence of pyridilethylation treat-
ment.
The amino acid analysis of the minor protein component X of A.
saxicola (AS) is also shown.
the amino acid analysis was carried out without prior pyr-
idilethylation, trace amounts of cysteine could still be de-
tected. Because of its relative electrophoretic mobility, X2
(see Fig. 1-AS) most probably represents the dimer form
of X. Indeed, X2 disappears upon /3-mercaptoethanol
treatment of the starting protein sample (see Fig. 5).
i
Figure 6. Analysis of the time course of digestion by trypsin of the
monomer (M) and dimer (D) of the PL-I protein ofAgriodesma saxicola.
Digestions were carried out in 2 A/ NaCl, 50 mA/ Na phosphate (pH
6.8), at an enzyme:substrate ratio of 1:500. The digestion times were: 0:
0 min; 1: 5 min; 2: 15 min; 3: 30 min; 4: 60 min and 5: 120 min. AS*:
nuclear sperm-specific proteins of A. saxicola slightly degraded by an
egg protease (see legend to Fig. 7). r: peptide resistant to digestion by egg
proteases.
Specific degradation oj PL-I in A. saxicola
Every specimen of .-1. saxicola analyzed was hermaph-
roditic. Although the male and female gonads are com-
pletely separated, some contamination of the sperm by
oocytes sometimes occurred when accidental incisions
were made in the ovary as the shells were being opened.
The protein composition of the crude HC1 nuclear extracts
thus obtained showed a complex and highly variable pat-
tern in urea acetic acid gel electrophoresis. Figure 7 shows
a light microscope and electrophoretic analysis of several
sperm samples with different amounts of contamination
by oocytes. In preparations containing pure sperm, the
electrophoretic analysis showed two major bands, M and
D (Fig. 7a). corresponding to the monomer and dimer of
PL-I, as well as a 15-20% of X and X2. As the extent of
contamination by female germinal cells increases (see Fig.
7b, c), the amounts of M and D present in the HC1 extracts
38
D
Figure 7. Microscopic analysis with phase contrast of pure sperm (A), and of sperm preparations containing
an increasing amount of contamination by eggs (B) and (C). The samples were obtained from different
specimens of Agrtiidesma saxicola. a, b, and c: electrophoretic analysis, in urea-acetic acid PAGE, of sperm
preparations containing an increasing amount of contamination by eggs. As contamination increases, an
extensive degradation of both the monomer (M) and dimer (D) forms of the major nuclear sperm-specific
PL-I component is observed. A relatively resistant peptide — r — is generated during this degradation process.
The white bar corresponds to 50 pm. X and X2 are as in Figure 5.
decrease, and the proteins finally disappear completely.
This is accompanied by the appearance of a complex pat-
tern of new bands with faster electrophoretic mobility (Fig.
7b, c). Such protein pattern transition is clearly indicative
of a degradation process elicited by specific proteases from
the contaminating eggs. A similar in vitro degradation of
sperm histones by the cytoplasm of sea urchin eggs has
also been reported (Betzalel and Moav, 1987). A quite
resistant degradation peptide. designated r, is produced
during this process. Although the composition and nature
of peptide r are completely unknown, it is certainly much
smaller than the trypsin-resistant peptide obtained under
in vitro conditions (see Fig. 6). A nuclear HC1 protein
extract from pure oocytes contained none of the proteins
observed in Figure 7 (results not shown).
Discussion
In this work I have analyzed the protein composition
of the nuclei of the sperm of two representatives of the
subclass Anomalodesmata within the class Bivalvia. In
both of the species analyzed — Agriodesma saxicola and
Mytilimeria nuttalli — 80-90% of the nuclear sperm-spe-
cific proteins consist of a mixture of dimer (D) and mono-
mer (M) forms of a protamine-like (PL-I) protein. The
remaining 10-20% includes the minor protein fractions
X and Y.
The protamine-like nature of the major proteins is re-
vealed by their amino acid composition (see Table I). They
clearly fulfill the compositional definition of protamines
(Subirana, 1983): (Lys + Arg) = 45-80%, (Ser + Thr)
= 10-25%. Indeed, of all the PL proteins characterized
so far, the ones analyzed here exhibit the highest arginine
content within the PL classification (Ausio, 1986). The
presence of a trypsin-resistant core (see Fig. 6) indicates
that these proteins are also related to the proteins of the
histone H 1 family. Therefore, the PL major components
of both A. saxicola and M. nuttalli should be undoubtedly
classified within the PL-I class (Ausio, 1986). The presence
of cysteine in these proteins does not seem to be an un-
usual feature; indeed, two cysteines also occur in the PL-
I component ofMacoma nasuta (Ausio, 1988).
Because the minor protein component X of A. saxicola
has an amino acid composition indistinguishable from
PL-I, these two proteins must be closely related, and X:
(see Fig. 1-AS) may represent a dimer form of X. The
same observations apply to the X and Y components of
M. nuttalli (results not shown). The structural relation-
ships among PL-I and the X and Y fractions remain ob-
scure, but the similarity of their amino acid analyses to
that of the major protein component suggests that X and
Y may be proteolytic peptides from PL-I. They could arise
from the activity of either a nuclear or an acrosomal sperm
protease (Miiller-Esterl and Fritz, 1981) during protein
extraction. However, they do not seem to be related to
any of the protein fragments produced by the protease
digestion resulting from oocytic contamination, because
NUCLEAR PROTEINS FROM THE SPERM OF BIVALVE MOLLUSKS
39
Table II
Classification of the class Bivalvia according to llieir protamine-like
group (Ansid. 1986)
Subclass (a)
Representative
species
Protein
type (b)
Reference
Cryptodonta
Palaeotaxodonta
Palaeoheterodon ta
Anodonta pisciniallis
[(?)
(c)
Heterodonta
Spisu/a solidissima
I
(d)
Ensis minor
I!
(e)
Macoma masuta
Ill
(f)
.\fytilus cilulis
IV
(g)
Pteriomorphia
Crassostrea gigas
0
(h)
Anomalodesmata
Agriodesma saxicola
II
(i)
Myli/imcria nuttalli
I
(i)
(a) According to Barnes (1980).
(b) According to Ausio (1986).
(c) Rozovrt al. (1984).
(d) Ausio and Subirana ( 1982a).
(e) Giancotti et al. (1983).
(I) Ausio (1988).
(g) Ausio and Subirana (1982c).
(h) Sellos(1985).
(i) This work.
the presence of X2 or X does not increase as the level of
egg-induced degradation increases (see Fig. 7). Indeed,
when whole sperm cells (without any prior preparation
of the nuclei) were extracted with HC1 for '/2 h at 4°C
immediately after sperm collection, the overall protein
pattern observed was undistinguishable from the pattern
of the HC1 extracts prepared from nuclei uncontaminated
by oocytes. In particular, the X-band was still observed.
The presence of a major PL-I protein in the sperm of
the two species analyzed here would allow us to classify
them within the protamine-like group I of my earlier clas-
sification of the bivalve mollusks (Ausio, 1986: see Table
II). Species fulfilling this compositional pattern have also
been described in other subclasses, including Palaeohet-
erodonta and Heterodonta. The presence of a PL-I pro-
tein, however, seems to be a common feature to all the
species of the class Bivalvia (Jutglar et al., 1991).
All of the PL-I proteins that have been analyzed in
detail have structures related to the histone H 1 superfam-
ily (Ausio e/ a/.. 1987; Ausio, 1988; Jutglar et al., 1991).
The structural similarities of PL-I to both histone H 1 and
the arginine-rich protamines from the vertebrates suggests
a close evolutionary relationship between these proteins.
In this sense, the increase in arginine and the decrease in
lysine and alanine observed in the case of the PL-I proteins
analyzed in the present work, when compared to other
PL-I proteins, would indicate a further departure from
their H 1 nature and a closer relationship to the protamines
from vertebrates.
Acknowledgments
I am very indebted to Debra Murie, Daryl Parkyn, and
Joachim Schnorr von Carosfeld from the Biology De-
partment at the University of Victoria for providing me
with the biological specimens used in this work. I am also
very grateful to Steve Carlos for his valuable assistance in
running the HPLC and FPLC columns and for reading
the manuscript. I also would like to thank Mrs. Denise
Lunger and Ms. Cheryl Gonnason for typing the manu-
script. This work was supported by NSERC Grant OGP
0046399 to Juan Ausio.
Literature Cited
Ando, T., M. Yamasaki, and K. Suzuki. 1973. P. 28 in Prolamines.
Vol. 12 of Molecular Biology, Biochemistry and Biophysics. Springer-
Verlag. Berlin. Heidelberg. NY.
Ausio, J., and J. A. Subirana. 1982a. A high molecular weight nuclear
basic protein from the bivalve molluscs. J. Biol. Chem. 257: 2802-
2805.
Ausio, J., and J. A. Subirana. 1982b. Conformational study and de-
termination of the molecular weight of highly charged basic proteins
by sedimentation equilibrium and gel electrophoresis. Biochemistry
21: 5910-5918.
Ausio, J., and J. A. Subirana. 1982c. Nuclear proteins and the orga-
nization of chromatin in spermatozoa of Mytilus edulis. Exp. Cell.
Res. 141: 39-45.
Ausio, J. 1986. Structural variability and compositional homology of
the protamine-like components of the sperm from the bivalve mol-
luscs. Comp. Biochem. Physiol. 85B: 439-449.
Ausio, J., A. Toumadje, R. McParland, R. R. Becker, W. C. Johnson,
Jr., and K. E. van Holde. 1987. Structural characterization of the
trypsin resistant core in the nuclear sperm-specific protein from Spi-
sula solidissima. Biochemistry 26: 975-982.
Ausio, J. 1988. An unusual cysteine-containing histone HI -like protein
and two protamine-like proteins are the major nuclear proteins of
the sperm of the bivalve mollusc: Macoma nasuta. J. Biol. Chem.
263: 10.141-10,150.
Barnes, R. D. 1980. Invertebrate Zoology. 4th ed. Saunders College,
Philadelphia.
Bernard, F. R. 1983. Catalogue of the living Bivalvia of the eastern
Pacific Ocean: Bering Strait to Cape Horn. Canadian Special Pub-
lication of Fisheries and Aquatic Sciences 81: 83.
Betzalel, M., and B. Moav. 1987. Degradation of sperm histones
in vitro by cytoplasm of the sea urchin egg. Cell Dtjier. 20: 125-
136.
Del.ange. 1976. Handbook of Biochemistry and Molecular Biology.
3rd ed. (Proteins), G. Fasman. ed. CRC press Boca Raton, FL, vol
II, 294 pp.
Giancotti, V., E. Russo, M. Casparini, D. Serrano, D. Del Piero, A. W.
Thorne, P. D. Car), and C. Crane-Robinson. 1983. Proteins from
the sperm of the bivalve mollusc: Ensis minor. Eur. J. Biochem 136:
509-516.
Hirs, C. M. \V. 1967. Determination of cysteine as cysteic acid. Me/hods
En:ymol. XI: 59-62.
Jutglar, L., J. I. Borell, and J. Ausio. 1991. Primary, secondary and
tertiary structure of the core of a histone HI -like protein from the
sperm of Mylilus. J Biol. Chem. 266: 8184-8191.
Kasinsky, H. E. 1989. Specifity and distribution of Sperm Basic Pro-
teins. Pp. 73-163 in Histones and Other Basic Nuclear Proteins. Vol.
I, L. S. Hnilica, G. S. Stein, and J. L. Stein, eds. CRC Press, Boca
Raton. FL.
40
J. AUSIO
Km hi. J. 1979. Synthesis of high mobility group proteins in regenerating
rat liver. / Biol. Chem. 254: 7276-7281.
Mogensen, C., S. Carlos, and J. Ausio. 1991 . Microheterogeneity and
interspecific variability of the nuclear sperm proteins from Mytilus.
FEES Lett. 282: 273-276.
Miiller-Esterl, W., and H. Fritz. 1981. Sperm acrosin. Methods En-
:ymol. 80: 621-632.
Odintsova, N. A., S. M. Rozov, and I. A. Zalenskaya. 1989. The chro-
mosomal proteins from the sperm of the bivalve molluscs Swiftopecten
swift v and Glycymerys yesonensis. Comp. Biochem. Physiol. 93B:
163-167.
Rozov, S. M., V. A. Brednikov, F. K. Corel, M. V. Lavrenteva, and
L. P. Solonenko. 1984. Structure of lysine-nch histone from sperm
of Anodonta piscinalis. Mol. Biol. (transl) (USSR) 18: 1497-1508.
Sellos, D. 1985. The histones isolated from the sperm of the oyster
Crassostrea gigas. Cell Differ. 17: 183-192.
Subirana, J. A., C. Cozcolluela, J. Palau, and M. Unzeta. 1973. Pro-
tamines and other basic proteins from spermatozoa of molluscs.
Biochim. Biophys. Ada 317: 364-379.
Subirana, J. A. 1983. Nuclear Proteins in Spermatozoa and Their In-
teractions with DNA. Pp. 197-214 in The Sperm Cell, J. Andre, ed.
Martinus Nijhoff B. V., The Hague.
Subirana, J. A., and J. Colom. 1987. Comparison of protamines from
freshwater and marine bivalve molluscs: evolutionary implications.
FEBS Lett. 220: 193-196.
Zalensky, A. O., and I. A. Zalenskaya. 1980. Basic chromosomal pro-
teins of marine invertebrates II. The proteins from the Bivalvia mol-
luscs. Comp. Biochem. Physiol. 66B: 415-419.
Reference: Biol Bull. 182: 41-53. (February, 1992)
The Origin of Cortical Vesicles and their Role in Egg
Envelope Formation in the "Spiny" Eggs of a
Calanoid Copepod, Centropages velificatus
PAMELA I. BLADES-ECKELBARGER' AND NANCY H. MARCUS2
1 Darling Marine Center, University of Maine, Walpole, Maine 04573 and Department of
Oceanography, Florida State University, Tallahassee, Florida 32306
Abstract. The mature oocytes of the marine calanoid
copepod, Centropages velificatus, contain two morpho-
logically distinct populations of cortical vesicles that un-
dergo sequential exocytoses at the time of spawning. The
contents of the primary cortical vesicles are released first
and form the primary egg envelope. This is followed by
the exocytosis of the secondary cortical vesicles. These
contain numerous intracisternal granules that, upon re-
lease into the perivitelline space, transform into a mass
of fine fibers. The continual accumulation of fibers con-
stitutes an extracellular matrix between the primary en-
velope and the egg's plasmalemma. Further amassment
of the fibers beneath the primary egg envelope results in
the formation of long, spiny projections. The evolution
of the cortical vesicles was traced to the early vitellogenic
oocytes and appears to be unique. The two populations
of cortical vesicles are synthesized together within the same
cisternal elements of rough endoplasmic reticulum (RER).
The RER originates from membranous blebs off both the
nuclear membrane and stacks of annulate lamellae in the
early vitellogenic oocytes. Numerous intracisternal gran-
ules are present within the RER. Some of these granules
fuse, forming a dense, ring-like structure in the extremities
of the cisternae. These bud off from the RER to become
the primary cortical vesicles. The unfused intracisternal
granules remain as discrete bodies within irregular profiles
Received 15 July 1 99 1 ; accepted 1 October 1991.
Contribution Nos.: (PBE) Harbor Branch Oceanographic Institution
Contribution No. 892 and Darling Marine Center, University ofMaine
Contribution No. 244; (NHM) Florida State University Marine Labo-
ratory Contribution No. 1068.
of vesicular ER and comprise the secondary cortical
vesicles.
Introduction
The subject of post-embryonic development in free-
living copepods has been a favorite research topic for de-
cades. Consequently, the literature abounds with descrip-
tions of naupliar and copepodid developmental stages.
However, studies relating to embyronic development, i.e.,
those stages from spawning to the emergence of the first
nauplius, are limited to only a few early publications
(Grobben, 1881; Fuchs, 1914; Witschi, 1934; Marshall
and Orr, 1954, 1955). In particular, details of the mech-
anisms of fertilization and egg envelope formation have
yet to be elucidated in the Copepoda.
Some marine calanoid species spawn their eggs into
ovisacs that remain attached to the female until the emer-
gence of the first or second naupliar stage. The majority
of marine calanoids, however, are broadcast spawners,
releasing the eggs freely into the surrounding water where
they undergo development. Within this latter group, eggs
of a variety of shapes and sizes and with different types
of surface ornamentation have been observed (Johnson,
1967; Koga, 1968; Kasaharac/a/.. 1974; Uye, 1983; Mar-
cus, 1990). The egg surfaces of most species are smooth,
but others may be adorned with flanges or spines of vary-
ing shapes and lengths. The production of spiny eggs has
been reported for numerous species, Acartia toma (Zil-
lioux, 1969), Centropages ponticus (Sazhina, 1968), C.
hamatus (Pertzova, 1974), Pontella mediterranea (Sa-
zhina, 1 968; Grice and Gibson, 1 98 1 ; Santella and lonora,
1990), A. erythraea, C yamadai, C. abdominalis (Kasa-
41
42
P. I. BLADES-ECK.ELBARGER AND N. H. MARCUS
hara et al, 1974), A. steuri (Uye, 1983), C. velificatus
(Marcus, pers. obs.), Calanus glacialis (J. Runge and
Blades-Eckelbarger, pers. obs.), and Candacia pachydac-
tyla (Blades-Eckelbarger, pers. obs.).
Some of the species listed above produce two morpho-
logical types of eggs where the spiny form represents a
diapause stage (hatching is delayed), and the smooth form
typifies a subitaneous stage (no mandatory delay in
hatching) e.g., Centropages hamatus (Pertzova, 1974) and
C. ponticits (Sazhina, 1968). Ponlella mediterranea pro-
duces three morphotypes; diapause eggs with long spines,
and subitaneous eggs that are either smooth or adorned
with short spines (Sazhina, 1 968; Grice and Gibson, 1981;
Santella and lonora, 1990). Acartiatonsa(ZH\iou\, 1969)
and A. steuri (Uye, 1983) have been reported to produce
both smooth and spiny eggs, but their physiological clas-
sification as diapause or subitaneous is still in question.
For the remaining species, only spiny eggs have been ob-
served, and there is no evidence to suggest that they are
a diapause stage.
While conducting a morphological survey of copepod
eggs found in sea bottom muds, we became intrigued with
the spiny modifications of the egg envelopes of some spe-
cies. Consequently, we initiated a study using light mi-
croscopy along with scanning and transmission electron
microscopy to investigate the stages of egg envelope de-
velopment and spine formation in the eggs of Centropages
velificatus.
Materials and Methods
Adult female Centropages velificatus (De Oliveira,
1947) were sorted from plankton tows collected approx-
imately 10 miles due east off the coast of Fort Pierce,
Florida. Female's carrying mature, pigmented oocytes
were placed in small dishes with filtered seawater that
were observed every few minutes for spawned eggs. The
eggs were carefully picked up by drawn-out pipettes and
placed onto pieces of 35 ^m mesh supported by Beem
capsules (Flood, 1973). The Beem capsules sat in shallow
glass dishes containing 2.5% glutaraldehyde in filtered
seawater.
For transmission and scanning electron microscopy
(TEM and SEM), eggs in varying stages of development,
from polar body extrusion to advanced spine formation,
were collected and fixed in this manner. After approxi-
mately 100 eggs were placed in a Beem capsule, the capsule
was transferred to a 5% Karnovsky's (1965) glutaralde-
hyde-paraformaldehyde mixture in 0.1 A/ Sorensens
phosphate buffer. The capsules were flushed several times
with the latter fixative to prevent precipitate caused by
seawater mixing with the phosphate buffer. As a matter
of convenience, due to the long duration of the complete
fixation process, the eggs were held in the Karnovsky's
fixative for varying times depending on the time of day
collected. Those collected in the morning were held at
room temperature for 3 to 6 h. Those collected in the
evening were held overnight at 4°C. The lower temper-
ature slows the fixation process. There were no apparent
differences in cell or organelle structure among the varying
times and temperatures.
Adult females carrying mature oocytes were prepared
also for TEM. Initially each individual was placed in a
small amount of the Karnovsky's glutaraldehyde mixture
for approximately 1 5 min. The head and urosome were
then removed with a sharp razor and the metasomes
transferred to a vial containing fresh fixative and held for
the same range of times as the eggs.
This primary fixation of both eggs and adult females
was followed by 2 or 3 rinses in 0. 1 Al Sorensen's phos-
phate buffer (pH 7.4) and then held in 2% OsO4 in 0. 1 M
Sorensen's buffer at room temperature for 1-2.5 h. The
samples were rinsed briefly with buffer and dehydrated
through an ascending series of alcohols to 70%. At this
point, some of the eggs were pipetted onto SEM stubs
covered with double-sided sticky tape and allowed to air
dry in a desiccator for 2 to 3 days. The air-dried stubs
were coated with gold palladium and observed with a Zeiss
Novascan 30 SEM.
For TEM examinations, the remaining eggs and female
metasomes were dehydrated further to 100% ETOH fol-
lowed by propylene oxide and infiltrated with three
changes of Epon (Luft, 1961). For final embedding, the
female metasomes were oriented in flat embedding molds.
The eggs were carefully drawn into a wide bore pipette
with fresh Epon and dropped into a Beem capsule, which
was centrifuged at room temperature for 20 min at setting
#6 in a clinical centrifuge. Because the Epon is of a slightly
thickened consistency, centrifugation is needed to con-
centrate the hardened eggs into the tip of the Beem cap-
sule. Extensive sectioning of eggs prepared in this manner
revealed no membrane or organelle damage. For light
microscopy, l-/um thick sections were cut with glass
knives on a Porter-Blum MT2B ultramicrotome, and
stained with Richardson's stain (Richardson et al,
1960). Thin sections for TEM were stained with uranyl
acetate followed by lead citrate and examined on a Zeiss
EM9-S2 TEM.
It should be noted that the procedures for both SEM
and TEM result in minor shrinkage of the eggs. Therefore,
all measurements are approximations.
Results
Live observations o/ spawning and spine formation
Females were observed spawning on several occasions,
during which they remained active, swimming in a normal
manner around the dish. The oocytes flowed out of one
COPEPOD EGG ENVELOPE FORMATION
43
Figures 1-5. SEMsof egg envelope and spine formation from emergence of first polar body (Fig. 1, unlabeled arrow) to 24-h-old embryo (Fig. 5).
Figure 6. SEM. High magnification of spines.
^m
• '"• • -'
t * x&:;3*R3
Figure 7. Perinuclear region of vitellogenic oocyte showing nuclear bleb (large arrowheads) extending from nuclear envelope (Nm) in formation
of rough endoplasmic reticulum (RER). Note intracisternal granules (g) within nuclear bleb and RER. NP, nuclear pores.
Figure 8. Perinuclear region of vitellogenic oocyte showing annulate lamellae (AL). Note swollen extremities (RER) containing granules (g). Nm.
nuclear membrane; Nu, nucleus.
44
COPEPOD EGG ENVELOPE FORMATION
45
or both oviducts, emerging from the genital pore as a single
mass. The female would periodically twitch the urosome,
causing the amorphous mass of eggs to break free and fall
to the bottom of the dish. Approximately 5-10 s after re-
lease from the female, the eggs separated from each other
and transformed from an oval to a spherical shape. Release
of the first polar body occurred at this time (Figs. 1, 14).
The second polar body was not observed. The actual pro-
cess of sperm and egg fusion in copepods has never been
reported, nor was it seen in the present study. Therefore,
it could not be ascertained when egg envelope formation
began relative to the moment of fertilization.
Figures 1 to 6 present comparative SEM views of the
stages of spine formation from emergence of the first polar
body (Fig. 1 ) to a 24-h-old embryo (Fig. 5). Approximately
5 min after spawning, large, rounded bumps appeared on
the egg surface (Fig. 2). These bulges became more slender
and pointed, forming short jagged spines (Fig. 3). It took
approximately 15-20 min for long spines to form. A sur-
vey of over 100 eggs that were at least 24 h old revealed
individual variations in the morphology, number, and
size of the spines.
Cortical vesicle formation in vitellogenic oocytes
Formation of the egg envelope in Centropages velifi-
catus involves the exocytosis of two morphologically dis-
tinct, membrane-bound inclusions present in the egg's
cytoplasm. Prior to spawning, the mature oocytes that
reside in the oviducts of the female contain a variety of
morphologically distinct granules, vesicles, and inclusions.
One type of inclusion, referred to here as the primary
cortical vesicle, appears as a membrane-bound body con-
taining an electron-dense, granular material that sur-
rounds an electron-lucent core (Figs. 11. 12). Favorable
sections through the center of these vesicles present the
appearance of a darkly staining ring around a flocculent
center (Fig. 12). The secondary cortical vesicles appear as
irregularly shaped vesicles filled with several moderately
dense granules (ca. 75-82 nm diameter) (Figs. 11, 12).
Primary and secondary cortical vesicles originate in the
very early stages of vitellogenesis, where a blebbing process
of the outer lamina of the nuclear membrane is observed
(Fig. 7). These nuclear blebs contain numerous moderately
dense granules (ca. 80 nm diameter) and pinch off to form
lamellar and vesicular profiles of rough endoplasmic retic-
ulum (RER). Stacks of annulate lamellae are also observed
in the perinuclear region (Fig. 8), as well as in the central
cytoplasmic region of mid- and late- vitellogenic stages (Fig.
9). Vesicles containing several dense granules, morpholog-
ically identical to the nuclear blebs, also pinch off from the
extremities of the annulate lamellae. Fusion of some, but
not all, of these intracisternal granules culminates in the
formation of the ring-shaped densities that characterize the
primary cortical vesicles (Figs. 8-10).
The cytoplasm of mid- vitellogenic oocytes is filled with
elongate profiles of RER containing numerous unfused,
intracisternal granules residing with one or more ring-
shaped densities (Figs. 9, 10). Small Golgi complexes were
observed infrequently, but did not appear to contribute
to the contents of the RER. In the mature oocytes, the
ring-shaped portions bud off from the RER to become
the primary cortical vesicles (Figs. 10, 11). They are en-
closed by a smooth membrane devoid of ribosomes. The
unfused intracisternal granules remain as discrete bodies
within irregular profiles of vesicular ER that also have lost
the attached ribosomes. These represent the secondary
cortical vesicles (Figs. 11, 12).
There is no elaboration of an egg envelope prior to
spawning. The oocytes are enclosed by a simple oolemma
that is coated with a lightly staining glycocalyx (Fig. 13).
The glycocalyx, or vitelline envelope, is deposited over
the oolemma by the associated follicle cells during the
mid-stages of vitellogenesis ( Blades- Eckelbarger and
Youngbluth, 1984).
Cortical reaction, egg envelope elaboration, and spine
formation
Deposition of the egg envelope results from a cortical
vesicle reaction involving two successive stages of exo-
cytosis. Soon after spawning, the majority of yolk bodies
and other inclusions accumulate toward the center of the
egg, but the primary and secondary cortical vesicles re-
main in the cortical cytoplasm (Fig. 14). The first cortical
reaction is characterized by the exocytosis of the primary
cortical vesicles. The bounding membrane of the primary
cortical vesicles fuses with the egg's plasmalemma, and
the enclosed material is released into the perivitelline space
(Figs. 15, 16). This results in the formation of a narrow
layer (ca. 20 nm thick) of darkly staining material situated
slightly above the egg's plasmalemma (Figs. 15, 16, 18,
2 1 ). We refer to this first layer as the "primary egg en-
velope." At this time, the egg surface has a "bumpy" ap-
pearance (Figs. 2. 3, 17) where regions of the plasmalemma
Figure 9. Early stage of primary cortical vesicle formation (large arrowheads) in vesicular RER of mid-vitellogenic oocyte. Note annulate lamellae
(AL) with swollen extremities (RER). M. mitochondrion; Y. yolk granules.
Figure 10. Mid-vitellogenic oocyte. Ring-shaped densities (large arrowheads) budding off of RER (*) in formation of primary' cortical vesicles.
M, mitochondrion; Y. yolk granule.
Figure 11. Late vitellogenic oocyle with primary cortical vesicles (Pv) now separate from secondary cortical vesicles (SV). M, mitochondrion.
Figure 12. High magnification showing structure of primary (Pv) and secondary (Sv) cortical vesicles.
Figure 13. Oolemma (Oi) of mature oocyte in oviduct of female covered by vitelline envelope (*). Fc. follicle cell; Oo, ooplasm.
Figure 14. Light micrograph, 1-^m-thick section of newly spawned egg and formation of first polar body (arrowhead). Note centrally located
yolk granules with primary and secondary cortical vesicles occupying cortical cytoplasm.
46
>Ffy;
•v
I ^ . * • •» - -V
'16 -
,-
..- "'
;
•
...
OPT--
E/T7*- ' "">- ii
" • I
19
- -7/ I m
•-* ~» '""' k """
**••
.. •
Figures 15 and 16. First cortical reaction. High magnification of egg surface showing fusion of primary cortical vesicles (Pv) with oolemma (Ol)
and exocytosis of dense material in formation of primary envelope (Pe).
Figure 17. Light micrograph. 1-fjm-thick section of egg during first cortical reaction and exocytosis of primary cortical vesicles. Note that
cytoplasm extends into projections of egg surface.
Figure 18. Cortical cytoplasm of egg in late stage of first cortical reaction showing fusion of granules within secondary cortical vesicles (Sv). M,
mitochondrion; Ol. oolemma; Pe. primary envelope; Y, yolk granule.
Figures 19 and 20. High magnification of secondary cortical vesicles showing fusion of granules.
47
48
P. I. BLADES-ECKELBARGER AND N. H. MARCUS
Pe
3frr-
-
H4*©l ** -^Sg"
:*^f£&fir ^Jf^LJL^^M *Z
" r -• ^^ir\w-^'<;
i
:.:.iTit"E*ii«?f?f^'^.'. / A'^ -^^>"'-^^r,'^/SL>.-..V.*fe'f.*---''*'i'. . ___________
Figure 21. Second cortical reaction showing exocytosis of granules (large arrowheads) from secondary cortical vesicles (Sv). Ol, oolemma; Pe,
primary envelope.
COPEPOD EGG ENVELOPE FORMATION
49
bulge out. The egg's cytoplasm projects into these ex-
panded areas (Fig. 17).
A second wave of exocytosis follows soon after the first
with the release of the intracisternal granules contained
within the secondary cortical vesicles (Figs. 21, 23). Just
prior to their release, however, some of the intracisternal
granules fuse with each other, forming slightly larger and
denser masses (Figs. 18-20). Once in the perivitelline
space, the intracisternal granules transform into a mesh-
work of fibers that adhere to the inner surface of the pri-
mary egg envelope (Figs. 22, 23, 25, 27, 28). Concomitant
with the second wave of exocytosis is the appearance of
numerous endocytotic pits and vesicles along the egg's
plasmalemma (Figs. 22, 23).
With the continual accumulation of fibers from the
secondary cortical vesicles, the primary envelope lifts
higher above the egg's plasmalemma forming an irregular
surface sculpturing (Fig. 22), the plasmalemma withdraws
from the core of the spines and the egg proper becomes
spherical again (Fig. 26). Observations of eggs in multi-
cellular stages (approx. 24 h), during or just after synthesis
of the naupliar cuticle, revealed both long and short spines
with a crenulated surface (Figs. 27, 28). The space between
the cuticle and the egg envelope is composed of a thick
mass of fibers (Figs. 27, 28).
Discussion
Based on our observations of the oocytes and eggs of
Centropages veliftcatus, we present here the first identi-
fication of cortical vesicles, and a description of the cortical
reaction and subsequent egg envelope formation in the
Copepoda. These processes follow the same general se-
quence of post-spawning events as reported in other an-
imal species (Schuel, 1985; Longo, 1988). However, where
the eggs of some species contain a single, morphologically
heterogeneous population of cortical vesicles (or granules),
those of C. velificatits were found to have two. Conse-
quently, the cortical reaction in the eggs of C. veliftcatus,
involves not one, but two exocytotic events.
The presence of structurally different populations of
cortical granules has been demonstrated in other crusta-
ceans, the crab Carcinus maemis (Goudeau and Becker,
1982) and the decapod shrimp Sicyonia ingentis (Pillai
and Clark, 1988, 1990). Talbot and Goudeau (1988) re-
ported four distinct cortical vesicles in the oocytes of the
lobster Homarus. In all cases, the various populations of
cortical vesicles exhibited distinctly different morpholo-
gies, underwent temporally separated exocytoses, and in
S1. ingentis (Pillai and Clark, 1990) were found to be
chemically heterogeneous. Each type of cortical vesicle
contributed to different layers of the egg envelope.
During the first cortical reaction in the eggs of Centro-
pages veliftcatus. the contents of the primary cortical ves-
icles form the outer, or primary, egg envelope. This layer
may correspond to the fertilization envelope of other an-
imals, which is formed from the mixing of the vitelline
layer with the exocytosed contents of the cortical vesicles
(Kay and Shapiro, 1985; Somers and Shapiro. 1989).
Exocytosis of the secondary cortical vesicles in the eggs
of C. veliftcatus follows soon after the primary egg enve-
lope is complete. The secondary vesicles contain several
discrete intracisternal granules that, upon their release into
the perivitelline space, transform into a myriad of fibers.
The accumulation of these fibers between the egg's plas-
malemma and the primary egg envelope forms an extra-
cellular matrix (ECM) that exhibits a similar morphology
to ECMs surrounding the eggs and embryos of other ma-
rine invertebrates (Spiegel el ai, 1989).
The present paper further illustrates the cellular mech-
anisms by which the two populations of cortical vesicles
are synthesized in the vitellogenic oocytes of Centropages
veliftcatus. In the oocytes of many animals, the cortical
vesicles are derived from the Golgi complex (see Schuel,
1985, and references therein). In the decapod shrimp, Si-
cyonia ingentis (Pillai and Clark, 1988), one population
of cortical vesicles is derived from Golgi complexes and
the second population from within the cisternae of RER.
Cortical vesicle formation in C. veliftcatus, in general, is
similar to that of Carcinus (Goudeau, 1 984) and Homarus
(Kessel, 1968: Talbot and Goudeau, 1988) in which the
vesicles are produced by the ER, and Golgi complexes do
not appear to contribute. Other aspects of cortical vesicle
formation in C. velificatus are unique; ( 1 ) both nuclear
blebs and annulate lamellae appear to be involved in for-
mation of the vesicular RER that synthesizes the intra-
cisternal granules and. (2) these intracisternal granules
appear to be the precursors of both the primary and sec-
ondary cortical vesicles. Fusion of some of these granules
within the RER cisternae forms the dense, ring-shaped
contents of the primary cortical vesicles. The other intra-
cisternal granules do not fuse, but remain distinct and
Figure 22. Second cortical reaction showing lifting of primary envelope (Pe) and filling of perivitelline space with fine fibers (*). Note surface
sculpturing of primary envelope as well as coated micropinocytotic pits (p) and vesicles (v) along the oolemma (Ol). Sg. granules from secondary
cortical vesicles.
Figure 23. Early spine formation showing massive exocytosis of secondary cortical vesicle (Sv). Large arrowhead denotes transformation of
granular material into fine fibers. P, coated micropinocytotic pits; Sg. granules from secondary cortical vesicles in perivitelline space.
Figure 24. Mid-stage of spine formation.
Figure 25. High magnification of penvitelline space in 24-h-old embryo showing transformation of granules (small arrowheads) from secondary
cortical vesicles into line fibers (*). Cu, early cuticle of nauplius.
Figure 26. Light micrograph of live 24-h-old embryo with advanced spine formation. Note that cytoplasm has receded from spines (small
arrowheads).
Figures 27 and 28. Advanced spine formation of 24-h-old embryo (Em) showing thick mass of fibers filling spines (*). Cu. cuticle of nauplius.
50
COPEPOD EGG ENVELOPE FORMATION
51
comprise the secondary cortical vesicles. The primary
vesicles separate from the secondary vesicles in the later
stages of vitellogenesis.
In the eggs ofCarcinus maenus (Goudeau and Lachaise.
1980a. b), the cortical vesicles of one type are filled with
"ring-shaped" granules that are the precursor of the main
layer of the embryonic capsule. The authors emphasized
that these ring-shaped granules are homologous to the
"disc-shaped granules" or "intracisternal granules" pre-
viously considered as endogenous yolk in the vitellogenic
oocytes of several decapod crustaceans ( Beams and Kessel,
1962, 1963: Kessel, 1968; Ganion and Kessel, 1972).
Subsequent studies have confirmed that, instead of pos-
sessing nutritive qualities, the ring-shaped granules in these
crustacean eggs play a structural role in formation of the
egg envelope (Goudeau and Becker, 1982; Goudeau, 1984;
Talbot and Goudeau, 1988; Pillai and Clark, 1990).
Within the Calanoida, the secondary cortical vesicles
of Centropages velificatus appear homologous to the "in-
tracisternal granules representing the endogenous yolk"
in the oocytes of Centropages typicus (Arnaud ct a/.. 1982)
and to the "granular form of type 1 yolk" in the oocytes
of Labidocera aestiva (Blades-Eckelbarger and Young-
bluth, 1984) and Ponlella mediterranea (Santella and la-
nora, 1990). Our present observations parallel those of
Goudeau and Lachaise (1980a, b) and Talbot and Gou-
deau (1988), illustrating that the intracisternal granules
previously assumed to represent endogenous yolk in co-
pepod eggs, are actually precursors of the egg envelope.
The distinctive morphology of the primary cortical vesicles
in C. velificatus, however, has no correlate in the eggs of
other copepod species studied thus far, even in the oocytes
of a congeneric species, C. typicus (Arnaud el a/.. 1982).
The fact that the eggs produced by C. typicus do not elab-
orate spines warrants a closer look at morphological dif-
ferences between the eggs of these congeners.
One consequence of the two exocytotic episodes in the
eggs of Centropages velificatus is the addition of large
quantities of membrane to the egg's plasmalemma. This
occurs when the limiting membrane of the cortical vesicles
fuses with the plasmalemma of the egg. However, the di-
ameter of the egg does not increase significantly. The
presence of numerous endocytotic pits and vesicles ob-
served along the egg's plasmalemma during the second
exocytotic event provides a mechanism for the recycling
of at least some of the extra surface membrane. This pro-
cess has been illustrated in the eggs of other animals (see
review by Longo, 1988) and conforms with similar ob-
servations on mammalian secretory tissues (Mata and
Christensen. 1990).
Earlier studies have described two membranes sur-
rounding the copepod egg (see reviews by Davis, 1968,
1981 ). During hatching, the outer membrane cracks and
the inner membrane pushes out. The outer membrane
slips off and the nauplius is enclosed within the more
delicate inner membrane. The nauplius then breaks open
this membrane with its appendages and swims free. The
ultrastructural features of the primary egg envelope in
Centropages velificatus do not exhibit a trilamellar com-
position indicative of a true membrane. Therefore, we
suggest that this layer should be referred to as the hatching
envelope, such as described for other crustaceans (Gou-
deau and Becker, 1982; Pillai and Clark. 1988). The pres-
ence and structure of an inner egg membrane around the
copepod nauplius has yet to be validated because we did
not examine the later embryonic stages.
The morphology of the subitaneous egg envelope of
Centropages velificatus is very different from that of the
envelope encasing diapause eggs as reported for Hemi-
diaptomus ingens privinciae (Champeau, 1970), Diapto-
mus sangiiineus (Hairston and Olds, 1984), Pontella
mediterranea (Santella and lanora, 1990), and Anomal-
ocera patersoni (lanora and Santella, 1991). The thick,
multi-layered envelope surrounding diapause eggs appears
as a lamellar arrangement of microfibrils in a helicoidal
array and is considered comparable to the typical integ-
ument of arthropods (Hairston and Olds, 1984).
The function of the spines that characterize the eggs of
Centropages velificatus and those of other copepod species
remains elusive. During a discussion session of the sym-
posium entitled "Cultivation of Marine Invertebrates"
held in Princeton in 1967, it was suggested that spines on
copepod eggs might retard sinking (Allen, 1969), enhance
gas exchange, and afford protection from predation (Shel-
bourne, 1969). However, while it seems reasonable that
the spines would deter predators, Zillioux ( 1 969) reported
that spiny eggs were consumed by adult female Acartia.
More recently, Santella and lanora (1990) suggested that
the four-layered egg envelope and accompanying spines
present on the diapause eggs of Pontella mediterranea may
supply nutritive material during diapause and provide ex-
tra protection from harsh environmental conditions.
Numerous functions have been proposed for the spines
that cover the surface of other marine invertebrate eggs.
The eggs of some sea urchins, starfish, and sea anemones
present a spiny appearance due to the elongation and
bundling of cortical microvilli (Schroeder, 1982, 1986).
It is suggested that these microvillous "spikes" and
"spires" play a role in reinforcing the egg surface (Schroe-
der, 1982), function to facilitate absorption, and aid in
adhesion between dividing cells of the embryo (Schroeder,
1986). Copepod oocytes are not known to possess mi-
crovillar modifications of the oolemma (Blades-Eckel-
barger and Youngbluth, 1984), nor do the eggs form mi-
crovilli after fertilization (present study). Furthermore, the
spines of Centropages velificatus are not cytoplasmic, but
are projections of the outer or primary egg envelope caused
by the amassment of an extracellular matrix (ECM) within
52
P. I. BLADES-ECKELBARGER AND N. H. MARCUS
the perivitelline space. Recent studies on the ECMs sur-
rounding the embryos of other marine invertebrates may
hint to the role of the ECM that coats the eggs of C. ve-
lificatus. ECMs are believed to provide support and pro-
tection to the developing cells, aid in cell movement and
cell adhesion, and form a semi-permeable filter for uptake
and concentration of substances from the environment
needed for growth and differentiation (Spiegel el a!., 1 989).
Literature Cited
Allen, J. A. 1969. Marine Biology Vol. V. Proceedings of the Fifth
Interdisciplinary Conference on Marine Biology, J. D. Costlow Jr.,
ed. Gordon and Breach, Science Publishers, New York, London, and
Paris. 606 pp.
Arnaud, J., M. Brunei, and J. Mazza. 1982. Etude de 1'ovogenese chez
Centropages typicus (Copepoda, Calanoida). Reprod. Nutr. Dev. 22:
537-555.
Beams, J. \V., and R. Kessel. 1962. Intracisternal granules of the en-
doplasmic reticulum in the crayfish oocyte. J. Cell Biol. 13: 158-
162.
Beams, J. W., and R. Kessel. 1963. Electron microscope studies on
developing crayfish oocytes with special reference to the origin of
yolk. J. Cell Biol. 18: 621-649.
Blades-Eckelbarger, P. I., and M. J. Youngbluth. 1984. The ultra-
structure of oogenesis and yolk formation in Labidocera aesliva (Co-
pepoda: Calanoida). / Morphol. 179: 33-46.
Champeau, A. 1970. Etude de la vie latent chez les Calanoides (Co-
pepodes). Caracteristiques des eaux temporaires de Basse-Provence.
Ann. Fac. Sci. Marseille 44: 155-189.
Davis, C. C. 1968. Mechanisms of hatching in aquatic invertebrate
eggs. Oceanogr. Mar. Biol. Ann. Rev 6: 325-376.
Davis, C. C. 1981. Mechanisms of hatching in aquatic invertebrate
eggs. II. Oceanogr. Mar Biol. Ann. Rev 19: 95-123.
De Oliveira, L. P. H. 1947. Estudos sobre o Microplancton capturado
durante a viagem do navio hidrografico Lahmeyer nas baias de Ilha
Grande e Sepetiba. Mem. Inst. Oswaldo Cm: 44: 441-488.
Flood, P. R. 1973. A simple technique for the prevention of loss or
damage to planktonic specimens during preparation for transmission
and scanning electron microscopy. Sarsia 54: 67-74.
Fuchs, K. 1914. Die KeimblatterentwickJung von Cyclops viridis. Zoo/.
Jalirh. Anal. 38: 103-56.
Ganion, L. R., and R. G. Kessel. 1972. Intracellular synthesis, transport,
and packaging of proteinaceous yolk in oocytes of Orconcctes im-
munis. J Cell Biol. 52: 420-437.
Goudeau, M. 1984. Fertilization in a crab: III. Cytodifferentiation of
vesicles enclosing ring-shaped elements involved in the cortical re-
action. Gamete Res. 9: 409-424.
Goudeau, M., and J. Becker. 1982. Fertilization in a crab. II. Cytological
aspects of the cortical reaction and fertilization envelope elaboration.
Tissue and Cell 14: 273-282.
Goudeau, M., and F. Lachaise. 1980a. Fine structure and secretion of
the capsule enclosing the embryo in a crab (Carcinus maenus (L.)).
Tissue and Cell 12: 287-308.
Goudeau, M., and F. Lachaise. 1980b. "Endogenous yolk" as the pre-
cursor of a possible fertilization envelope in a crab (Carcinus maenas).
Tissue Cell 12:503-512.
Grice, G. D., and V. R. Gibson. 1981. Hatching of eggs of Pontella
medilerranea Claus (Copepoda: Calanoida). I'ie et Milieu 31: 49-
51.
Grobben, C. 1881. Die Entwicklungsgeschichte von Cetochilus sep-
temtrionalis Goodsir. Arb. Zoo/. Inst. Univ. H'ien 3: 1-40.
Hairston, N. G., and E. J. Olds. 1984. Population differences in the
timing of diapause: adaptation in a spatially heterogeneous environ-
ment. Oecologia 61: 42-48.
lanora. A., and L. Santella. 1991. Diapause embryos in the neustonic
copepod Anomalocera patersoni. Mar. Biol. 108: 387-394.
Johnson, M. W. 1967. Some observations on the hatching of Tortanus
discaudatus eggs subjected to low temperatures. Limnol. Oceanogr.
12:405-410.
Karnovsky, M. J. 1965. A formaldehyde-glutaraldehye fixative of
high osmolarity for use in electron microscopy. /. Cell Biol 27:
137A.
Kasahara, S., S. Uye, and T. Onbe. 1974. Calanoid copepod eggs in
sea-bottom muds. Mar. Biol 26: 167-171.
Kay, E. S., and B. M. Shapiro. 1985. The formation of the fertil-
ization membrane of the sea urchin egg. Pp. 45-80 in Biology of
Fertilization. A. Monroy and C. B. Metz. eds. Academic Press,
Inc., New York.
Kessel, R. G. 1968. Mechanisms of protein yolk synthesis and depo-
sition in crustacean oocytes. Z Zellforsch. 89: 1 7-38.
Koga, G. 1968. On the pelagic eggs of Copepoda. J. Oceanogr. Soc.
Jpn. 24: 16-20.
Longo, F. J. 1988. Reorganization of the egg surface at fertilization.
Int. Rev Cylol. 113: 233-269.
Luft, J. H. 1961. Improvements in epoxy resin embedding methods.
/ Biophys. Biochem. Cytol. 9: 409-414.
Marcus, N. H. 1990. Calanoid copepod, cladoceran, and rotifer eggs
in sea-bottom sediments of northern California coastal waters:
identification, occurrence, and hatching. Mar. Biol. 105: 413-
418.
Marshall, S. M., and A. P. Orr. 1954. Hatching in Calanus fmmar-
chicus and some other copepods. J. Mar. Biol. Assoc. U.K. 33: 393-
401.
Marshall, S. M., and A. P. Orr. 1955. The biology of a marine copepod,
Calanus finmarchicus (Gunnerus). Oliver & Boyd, Edinburgh and
London.
Mata, R. L., and E. I. Christensen. 1990. Redistribution and recycling
of internalized membrane in seminal vesicle secretory cells. Biol.
Cell. 68: 183-193.
Pertzova, N. M. 1974. Life cycle and ecology of a thermophilous co-
pepod, Centropages hamalus in the White Sea. Z Zhurnal 53: 101 3-
1022.
Pillai, M. C., and W. H. Clark Jr. 1988. Hatching envelope formation
in shrimp (Sicyonia ingentis) ova: origin and sequential exocylosis
of cortical vesicles. Tissue Cell 20: 941-952.
Pillai, M. C., and W. H. Clark Jr. 1990. Development of cortical ves-
icles in Sicyonia ingentis: their heterogeneity and role in elaboration
of the hatching envelope. Mol. Reprod. Dev 26: 78-89.
Richardson, K. C., L. Jarett, and E. H. Finke. 1960. Embedding in
epoxy resins for ultrathin sectioning in electron microscopy. Stain
Technol. 35: 313-323.
Santella, L., and A. lanora. 1990. Subitaneous and diapause eggs in
Mediterranean populations of Pomella medilerranea (Copepoda:
Calanoida): a morphological study. Mar. Biol 105: 83-90.
Sazhina, L. I. 1968. On hibernating eggs of marine Calanoida. Zoo/.
Zhurnal 47: 1554-1556.
Schroeder, T. E. 1982. Novel surface specialization on a sea anemone
egg: "spires" of actin-filled microvilli. J. Morphol. 174: 207-216.
Schroeder, T. E. 1986. The egg cortex in early development of sea
urchins and starfish. Pp. 59-100 in Developmental Biology. L. W.
Browder, ed. Plenum Publishing Corp.
Schuel, H. 1985. Functions of egg cortical granules. Pp. 1-43 in Biology
of Fertilization. A. Monroy and C. B. Metz, eds. Academic Press,
Inc.. New York.
COPEPOD EGG ENVELOPE FORMATION
53
Shelbourne, J. C. 1969. In Marine Biology Vol. V Proceedings of the
Fifth Interdisciplinary Conference on Marine Biology. J. D. Costlow
Jr., ed. Gordon and Breach, Science Publishers, New York, London,
and Paris. 606 pp.
Somers, C. E., and B. M. Shapiro. 1989. Insights into the molecular
mechanisms involved in sea urchin fertilization envelope assembly.
Dev. Growth Differ. 31: 1-7.
Spiegel, E., L. Howard, and M. Spiegel. 1989. Extracellular matrix
of sea urchin and other marine invertebrates. J. Morphol. 199:
71-92.
Talbot, P., and M. Goudeau. 1988. A complex cortical reaction leads
to formation of the fertilization envelope in the lobster, Homarus.
Gamete Res. 19: 1-18.
Uye, S. 1983. Seasonal cycle in abundance of resting eggs of Acartia
steuri Smirnov (Copepoda, Calanoida) in sea-bottom mud of Ona-
gawa Bay, Japan. Criistaceana 44: 103-105.
Witschi, E. 1934. On determinative cleavage and yolk formation in
the harpactid copepod Tisbe jurcata (Baird). Bio/. Bull 68: 335-340.
Zillioux, E. J. 1969. In Marine Biology Vol. V, Proceedings of the Fifth
Interdisciplinary Conference on Marine Biology, J. D. Costlow Jr.,
ed. Gordon and Breach, Science Publishers, New York, London, and
Paris. 606 pp.
Reference: Biol. Bull. 182: 54-65. (February. 1992)
The Role of Shell Granules and Accessory Cells
in Eggshell Formation in Convoluta pulchra
(Turbellaria, Acoela)
RESA M. CHANDLER,1 MARY BETH THOMAS,1 AND JULIAN P. S. SMITH, III2
Department of Biology, The University oj North Carolina at Charlotte,
Charlotte, North Carolina 28223
Abstract. Most turbellarian embryos are surrounded by
a sclerotinized eggshell originating from polyphenol-con-
taining eggshell-forming granules (EFGs). Although em-
bryos of the acoel Convoluta pulchra are surrounded by
a shell, it is not sclerotinized. Therefore, in the absence
of polyphenols as a marker for EFGs, it was not clear
which, if any, of the granules of the oocyte function in
eggshell synthesis. In this study, electron-opaque, elliptical
granules with a characteristic frothy component and a
diameter of 480 nm were identified in the oocyte as EFGs
by their participation in eggshell formation. In addition,
it was shown that accessory cells to the oocyte initiate
synthesis of the shell by producing a thin, granular, elec-
tron-opaque primary shell, against which the contents of
the EFGs are released by exocytosis. Morphological com-
ponents of the shell and stages of its synthesis are de-
scribed. A second type of membrane-bound granule and
the lipid droplets that occur in the ooplasm were found
not to be involved in eggshell formation and are probable
sources of nutrients for the developing embryo. Possible
implications of the findings for taxonomy and phylogeny
are discussed.
Introduction
The zygotes of acoel turbellarians, like those of all pla-
tyhelminthes studied to date, are enclosed in a shell fol-
lowing fertilization (see Rieger et ai, 1991, for Turbellaria;
Fried and Haseeb. 199 1 , and Coil, 199 1 . for parasitic pla-
tyhelminths). In all cases so far examined, the eggshell
Received 24 May 1991; accepted 31 October 1991.
Present addresses: 'Department of Biology, The University of North
Carolina at Charlotte. Charlotte, NC 28223, and :Department of Biology.
Winthrop College, Rock Hill. SC 29733.
appears to arise from eggshell-forming granules (EFGs)
exocytosed from one of the cells that ultimately comes to
lie within the shell (oocytes in the entolecithal archo-
ophorans, yolk cells in the ectolecithal neoophorans and
parasitic platyhelminths). Among archoophoran turbel-
larians, EFGs have been described by transmission elec-
tron microscopy (TEM) from the oocytes of at least one
member of most orders (Polycladida: Boyer, 1972; Do-
menici et ai, 1975; Gammon, 1979; Ishida et ai, 1981;
Espinosa, 1986; Ishida and Teshirogi, 1986; Macrosto-
mida: Gremigni et al.. 1987; Kucera, 1987; Acoela: Gre-
migni, 1988; Smith et al., 1 988; Nemertodermatida: Smith
et ai, 1988). Relatively few TEM studies of archoophor-
ans, however, have examined formation of the eggshell
(see Ishida, 1989; Falleni and Gremigni, 1989).
Whereas EFGs in most platyhelminths examined to
date contain polyphenols, they have not been found in
the oocytes of the Acoela (Thomas et ai, 1985; Chandler
and Thomas, 1986, 1987; Gremigni, 1988; Smith et ai,
1988; Falleni and Gremigni, 1989, 1990) or the Nemer-
todermatida (Thomas et ai. 1985; Smith et ai. 1988),
the presumably primitive turbellarian orders that consti-
tute the Acoelomorpha. In the acoels, the shell is protein-
aceous and non-sclerotinized (Falleni and Gremigni,
1989). Because the eggshells of all other non-acoelomor-
phan platyhelminths studied so far appear to be sclero-
tinized, the process of eggshell formation in the acoels
merits further study. For example, in the absence of the
polyphenolic marker, it is difficult to ascertain with cer-
tainty which, if any, of the several types of granules in the
oocyte give rise to the eggshell. Falleni and Gremigni
(1989) have implicated a population of electron-opaque
granules with a diameter of 0.4-0.5 /urn as EFGs in the
oocytes of the acoel " Convoluta psammophy la" (? = Pae-
54
SHELL DEPOSITION IN AN ACOEL
55
domecynostomum psammophilum Beklemischev, 1957),
but have not examined the mechanism by which these
granules contribute to the formation of the shell. Although
it seems likely that exocytosis of the granules, which occurs
during eggshell formation in other turbellarians, is in-
volved in production of the eggshell of the acoels (Falleni
and Gremigni, 1990), that is not clear from the published
studies.
Also unclear is the question of whether cells other than
the oocyte are involved in synthesis of the eggshell in
acoels, as they are in other turbellarians (e.g., Giesa, 1966;
Bunke, 1982; Ishida, 1989). Mature oocytes in both acoels
and nemertodermatids are nearly always surrounded by
"follicle" or "accessory" cells, whose function has not
been elucidated, although it is usually suggested that they
are responsible for heterosynthetic yolk production or that
they assist in the production of the eggshell (see Rieger et
ai. 1991).
The present study examines eggshell formation in the
acoel Convoluta pulchra with particular attention to the
following questions: ( 1 ) do any of the granules of the oo-
cyte participate in the formation of the eggshell; (2) if so,
what is the mechanism by which they participate; (3) are
other cell types involved in eggshell synthesis; and (4) what
is the morphology of the shell itself? A preliminary report
of these findings has been presented elsewhere (Chandler
eta/.. 1988).
Materials and Methods
Experimental organism
Convoluta pulchra (Family Convolutidae; Smith and
Bush, 1991) was extracted according to the methods of
Hulings and Gray ( 1 974) from sediment samples collected
at mid-tide level at a coastal inlet near Fort Fisher, North
Carolina.
Procedures for preliminary observations
To determine the time course of egg-laying in Con-
voluta pulchra. gravid acoels were isolated in pairs in
wells of Falcon® 96-well plates. The individual cultures
were maintained in Millipore®-filtered seawater (MFSW)
at a constant temperature of 20°C and a light:dark cycle
of 16:8 h. The worms were monitored closely to deter-
mine if the acoels lay their eggs according to a diurnal
pattern. Under these conditions, worms began egg-lay-
ing within approximately 1 to 1.5 h after exposure to
light.
Experimental design
Ten pairs of acoels with large eggs were placed in Fal-
con® plates. The animals were placed into darkness at
9:30 p.m. and returned to light at 6:30 a.m. Worms were
fixed for electron microscopy every half hour, from 7:00
a.m. until 1 1:30 a.m. Other worms were allowed to lay
eggs, which were fixed for electron microscopy.
Procedures for electron microscopy
Adult worms and laid eggs were fixed in 1% glutaral-
dehyde, 4% paraformaldehyde, 0. 1 M HEPES buffer (pH
7.4), 1 mA/ CaCl2, and 10% sucrose [modified from
McDowell and Trump (1976)], rinsed in buffer, post-fixed
in HEPES-buffered 1% OsO4, and embedded in Spurr's
low viscosity epoxy resin (Spurr, 1969). Ultrathin sections
were stained with uranyl acetate (Watson, 1958) and lead
citrate (Reynolds, 1963), and examined with a Philips
20 1C transmission electron microscope.
Procedures for morphometric analysis
The Feret diameters (see Weibel, 1979) of the three
types of granules were measured with a Zeiss ZIDAS dig-
itizer. The Feret diameter of each granule profile was
measured between two lines parallel to the long axis of
the photographic print. The most mature stage of devel-
opment in which all types of granules were present (Pri-
mary Shell Synthesis Stage, described below) was chosen
for measurement. To avoid measuring the same granule
more than once, 1 .8 /nm separated the thin sections mea-
sured and non-overlapping micrographs were taken from
each thin section examined. Approximately 200 granules
from electron micrographs magnified 28.000X were mea-
sured and the size-frequency distribution of Feret diam-
eters for granule Types A and B were plotted. For the
Type A granules, the distribution was corrected for profiles
overlooked in the smallest categories as described in Wei-
bel (1979). The actual diameter of Type A granules (D)
was estimated from the average Feret diameter (d) using
4
the relationship D ^ --d (Weibel, 1979).
7T
Oocytes at different stages of shell maturation were an-
alyzed to determine the volume densities ( Vv; % of oocyte
volume occupied by granules) of Types A and B granules
and lipid droplets. Non-overlapping micrographs along
two right-angled transects were taken from the germinal
vesicle to the oolemma. Volume densities were deter-
mined by point-count analysis, using the oocyte as the
reference space (Weibel. 1979). To determine whether
these volume densities changed in the oocyte as devel-
opment of the shell proceeded, the volume densities were
arcsin-square root transformed and subjected to ANOVA
Planned Comparison.
The embryos contained within laid eggs fixed for elec-
tron microscopy were observed to be separated from the
inner edge of their shells. This could occur if the shell
swells and lifts away from the embryo or if the embryo
•'•., •*••• - *u*-;
> <4 '
D» _ .._.- » i •»
» • f i j ^ •*
. , .,, -
1 « c • * • - *
:. >
Figure \. (A) Overview of a column of maturing oocytes showing the nuclei (germinal vesicles) with
large nucleoli (Nu) and the increase in the number ot granules in the ooplasm with development. Ol
= oocyte synthesizing Type A granules only: O2 = oocyte synthesizing Types A and B granules; O3 = oocyte
containing Type A and B granules and lipid droplets. Bar = 15.0 ^m. (B) Cisternae of rough endoplasmic
56
SHELL DEPOSITION IN AN ACOEL
57
Type A
16
5 6 7 8 9 1C 11
Feret Diameter (x 100 nm)
13 H 15
TypeB
find
_n
567 8 9 1C 11
Feret Diameter (x 100 nm)
12 13 14 15
Figure 2. Feret diameter distribution of granule Types A and B.
Each number on the abscissa represents the stated value ±0.5; n = 200.
loses some of its volume, shrinking away from the shell.
Since the two possibilities affect the interpretation of the
changes in volume densities of the granules, the absolute
volumes of eggs prior to egg-laying and of embryos after
egg-laying were determined. Serial 2 ^in-thick sections
were viewed with a Zeiss Axioskop light microscope
equipped with a Sony DXC-3000A color video camera
and a Sony PVM-1343MD Trinitron color monitor. Each
section of egg or embryo was traced from the monitor
screen onto transparent plastic. The ZIDAS was used to
calculate the area of each drawn section; the area, in turn,
was multiplied by the thickness of the sections and these
numbers summed for all sections to determine the volume
of each egg or embryo. Four eggs with mature shells and
four laid eggs were analyzed. ANOVA was used to com-
pare the mean volume of eggs with mature shells to that
of embryos within the laid eggs.
Results
Stages of shell formation
Based on the ultrastructural features described in detail
below, development of the eggshell in Convoluta pulchra
-. 14 -
10 -
Ol
D
m
6
3
I
• Type A
E3 TypeB
D Lipid
JL
PS
M
Stage of Shell Synthesis
Figure 3. Volume densities of granular inclusions. PS = stage of
Primary Shell synthesis; MS = stage of Mature Shell synthesis; M
= stage of Mature Shell; L = stage of Laid Shell.
was divided into four sequential stages: Primary Shell
Synthesis, Mature Shell Synthesis, Mature Shell, and Laid
Shell. The germinal vesicle persists into the Mature Shell
stage, during which it breaks down. Because the time of
fertilization is not known for the species, the Mature Shell
stage may be an oocyte or a zygote, and will therefore be
referred to as the "egg" (unfertilized or fertilized). At the
Laid Shell stage, embryos are developing.
Description of granules and lipid droplets
Granules we have termed Type A granules were the
first to appear in developing oocytes (Fig. 1 A, egg 1 ). Oo-
cytes at this stage had a large germinal vesicle containing
euchromatin and a single prominent nucleolus. Long
strands of rER occurred in the cytoplasm. Additional rER
and Golgi were often found close to one another, usually
with the rER almost encircling the Golgi in a horseshoe
configuration (Fig. 1 B). Mitochondria were abundant. The
forming Type A granules were spherical to ellipsoidal and
electron-opaque with slightly lucent internal areas (Fig.
1C); profiles of the granules did not exceed 330 nm in
early oocytes.
In more mature primary oocytes, the Type A granules
were more complex, larger, and more abundant. Mature
Type A granules consisted of two components, electron-
reticulum (rER) almost totally encircling the Golgi (G) in an oocyte synthesizing Types A and B granules.
Bar = 0.5 Mm. (C) An immature Type A granule before the frothy element is evident. Only the electron-
opaque component (E) is present. The membrane surrounding the granule is indicated by the arrow. Bar
= 0.5 Mm. (D) Two mature Type A granules with their typically elliptical profile. Both the electron-opaque
component (E) and the frothy element (F) can be distinguished. Bar = 0.5 Mm. (E) A field of granules from
an oocyte slightly older than oocyte 3 in Figure 1. Type A granules (A), Type B granules (B). and lipid
droplets (L) occur in the cytoplasm. Bar = 1 .0 /im.
r
AC
^^^^
4B
-• - . - * ^?fe.\<^«^>?>yf'^s71
" " -
•' "
- :'-
O
I
. <
AC
Figure 4. (A) An oocyte (O) with a primary shell (arrows). Type A granules (A). Type B granules (B),
and lipid droplets (L) can be identified in the ooplasm. Bar = 5.0 /urn. (B) A region of the surface of an
oocyte (O) in contact with an accessory cell (AC). In the pan of the region of contact indicated by the arrows
the granular primary shell has formed. Bar = 1.0 iim. (C) An accessory cell (AC) with an extensive rough
40
58
SHELL DEPOSITION IN AN ACOEL
59
opaque globules and a frothy material that often formed
a cap at one pole of the granule (Fig. 1 D, 1 E). The mature
Type A granules from an oocyte in the Primary Shell
Synthesis stage exhibited an average Feret diameter of
380 nm (n = 199), giving an estimated average diameter
of 480 nm; the largest profile measurement was 640 nm
(Fig. 2). Type A granules reached an average volume den-
sity of 2.9% in Primary Shell Synthesis stage oocytes (Fig.
3). Although Type A granules were not observed to mar-
ginate at the time of eggshell formation, their volume
density dropped to 1 .4% in the single Mature Shell Syn-
thesis stage oocyte observed, and was zero in both Mature
Shell and Laid Shell stages (Fig. 3).
A second type of granule, termed the Type B granule,
was first observed in oocytes that still possessed a large
germinal vesicle (Fig. 1A, egg 2). More Golgi bodies and
rER occurred in these oocytes than in those synthesizing
only Type A granules. Mature Type B granules were ir-
regular spheres with a flattened edge (Fig. IE). The Type
B granules were less electron-opaque than the Type A
granules and contained internal electron-lucent patches.
The mature Type B granules from a Primary Shell Syn-
thesis stage oocyte exhibited an average Feret diameter of
590 nm (Fig. 2); the largest measurement was 1570 nm
(n = 200). Calculation of an average diameter from the
Feret diameter was not attempted because Type B granules
appeared to depart significantly from a spherical shape.
The volume density of Type B granules was 5.5% in Pri-
mary Shell Synthesis stage oocytes as well as in the single
Mature Shell Synthesis stage oocyte, 5.9% in Mature Shell
stage oocytes, and 1 1 .6% in Laid Shell stage embryos (Fig.
3). The volume densities of Type B granules were not
statistically different during shell deposition (comparing
Primary Shell Synthesis stage oocytes to Mature Shell stage
oocytes: F = 0.004, df = 1, 10); however, the increase
seen in laid eggs was significant (F == 1 1.08, df = 1, 10;
P < .05).
Lipid droplets (Fig. IE) appeared in oocytes that had
begun synthesis of Type B granules, but were still in the
germinal vesicle stage (Fig. 1 A, egg 3). Lipid droplets ap-
peared to have no membrane. The volume density of lipid
droplets was 2.7% in Primary Shell Synthesis stage oocytes,
1.8% in the single Mature Shell Synthesis stage oocyte,
3.9% in Mature Shell stage oocytes, and 7.0% in laid eggs
(Fig. 3). As was the case for the Type B granules, the
volume density of lipid droplets was statistically constant
during shell synthesis (F = 0.22, df= 1, 10), but increased
significantly in laid eggs (F = 9.13, df = 1, 10; P < .05).
Morphology of eggshell deposition
Primary Shell Synthesis stage. Synthesis of the shell
began before the germinal vesicle had broken down. At
this stage, both Type A and Type B granules were dis-
persed throughout the ooplasm, and the oocyte was sur-
rounded wholly or in part by an accessory cell. Electron-
opaque material of finely granular composition appeared
outside the oocyte, along the irregular contours of the
oolemma (Figs. 4A, 4B). This layer was discontinuous in
the youngest oocytes of this stage examined, and was only
observed where the accessory follicle cell, laden with rER
(Figs. 4C, 4D), was in contact with the oolemma. In
slightly later stages, this thin primary shell covered the
entire oocyte as a layer approximately 50 nm thick.
Mature Shell Synthesis stage. In the single oocyte of
this stage encountered, numerous examples of exocytosis
of Type A granules were observed (Fig. 5A). Both the
homogeneous electron-opaque and the frothy compo-
nents appeared to be extruded from the cell. Fused gran-
ules often produced elongated channels or sacs, the mem-
branes of which were continuous with the plasmalemma
(Fig. 5 A).
Mature Shell stage. Released Type A granules appar-
ently produced a homogeneous, electron-opaque layer
100-525 nm thick immediately underneath the granular,
exogenously originated primary shell, as well as an inner
fibrillar network with a thickness of 10-315 nm (Fig. 5B).
Clear peripheral vesicles, presumably the remnants of the
Type A granules, were visible shortly after exocytosis (Fig.
5B). During the process of exocytosis of the Type A gran-
ules, small membrane-bounded fragments of cortical cy-
toplasm appeared to have been lost from the oocyte (Figs.
5A, 5B).
In somewhat older eggs, a new layer of the shell was
observed (Fig. 5C). This thin layer was sandwiched be-
tween the granular primary shell and the homogeneous
layer produced by the Type A granules and appeared as
a stripe that was more electron-opaque than the homo-
geneous layer. The fibrillar layer was still present just out-
side the plasmalemma. At this stage, no Type A granules
were apparent in the cytoplasm (Fig. 5D).
Laid Shell stage. Laid eggs were encapsulated in a clear,
flexible, yet sturdy shell (Fig. 6A). In the laid eggshell, the
endoplasmic reticulum (rER) in contact with an oocyte (O). Arrows indicate the primary shell that has
formed along part of the region of contact between the accessory cell and the oocyte. A Type A granule (A)
lies close to the region of formation of the thin shell, but exocytosis has not begun at this stage. Bar = 0.5
fim. (D) A region of the plasmalemma of an oocyte (O) with (arrows) and without (darts) an adjacent
accessory cell (AC). Note that the primary shell (arrows) is present only where there is an accessory cell. Bar
= 0.5 M"i.
* . i
01
5D
Figure 5. (A) Exocytosis of Type A granules. The contents of Type A granules (Al) are released from
the more mature oocyte (Ol) to form the layers of its mature shell (S) interior to the thin primary shell
(TS1). Darts indicate the long, often tortuous profiles of the membranes of the Type A granules during
exocytosis. The asterisk marks a small membrane-hounded fragment of cytoplasm that may no longer be
60
SHELL DEPOSITION IN AN ACOEL
61
fibrillar layer could be distinguished, but the granular pri-
mary shell, the homogeneous layer, and the stripe could
no longer be differentiated: instead, a single, thick, elec-
tron-opaque layer occurred peripheral to the fibrillar layer
(Fig. 6B). The thickness of the shell was far more ho-
mogeneous than that of the forming shell (Fig. 6B). There
was a space between the embryo and the shell (Fig. 6A-
C). The juvenile worm hatched one day after egg-laying.
Before hatching, the worm moved vigorously until the
shell gave way.
The mean volume of a Mature Shell stage egg was
325 X 103 ^m3 (n = 4), whereas the mean volume of
an embryo after deposition was 180 X 103 ^m3 (n = 4).
The volume of the embryo was therefore significantly
smaller than that of the unlaid egg (F = 33.6, df = 1,
6: P< 0.005).
Discussion
Eggshell synthesis
Our study demonstrates that the first element of the
eggshell in Convoluta pitlchra is deposited by accessory
cells before the Type A granules undergo exocytosis. Ac-
cessory cells, laden with rER, appear to envelop an oocyte
at the time of primary shell synthesis. This is evidenced
by the observation that the thin, granular, primary shell
layer appears outside the oolemma in areas of the oocyte
surface abutting an accessory cell, whereas regions that
do not abut an accessory cell are not covered by primary
shell. A role in production of the shell was previously
hypothesized for the accessory cell, but not demonstrated
(Falleni and Gremigni, 1990). The mechanism by which
the material that composes the thin shell is released from
the accessory cell is not known. Exocytotic vesicles have
not been identified. It is possible that vesicles simply have
not been detected, perhaps because they are small, present
in low numbers, or do not accumulate in the cytoplasm
of the accessory cell. Alternatively, the primary shell may
be a product of a reaction between substances located on
the surface of the surrounding accessory cell and on the
surface of the oocyte. Clearly, the material appears only
where the two cells are in contact.
Functionally, the Type A granules of C. pulchra are
comparable to EFGs of other turbellarians. They undergo
exocytosis at the Mature Shell Synthesis stage and are
absent from the egg at the end of the Mature Shell stage,
indicating that they participate in eggshell deposition.
After the formation of the primary shell, the contents
of the Type A granules released by exocytosis become
packed against the thin shell made earlier, eventually
forming a homogenous layer beneath the primary shell
layer. This homogeneous layer probably comes from the
electron-opaque portion of the Type A granules. The ma-
ture shell enclosing the unlaid egg is characterized by a
fibrillar network that forms the innermost layer. The floc-
culent portion of the Type A granules most likely produces
this fibrillar network. These conclusions are based on
morphological observations; the detailed cytochemistry
of Type A granules is unknown. Also characteristic of the
mature shell is an electron-opaque stripe between the pri-
mary shell layer and the homogeneous layer. This could
represent a zone of reaction between the granular primary
shell layer formed by the accessory cell and the homo-
geneous layer formed from the Type A granule, perhaps
associated with some as yet unidentified hardening pro-
cess. This is suggested by the observation that in the shell
of the laid egg, the three outermost layers are no longer
discrete. This would be predicted if a reaction between
the primary shell and the homogeneous layer produced
the stripe and that reaction proceeded until, in the laid
eggshell, the stripe replaced the two original layers.
Although the wide variation in the thickness of the
components of the shell during early stages of synthesis
may be related in part to plane of section, it more likely
reflects the number of Type A granules released in a given
area. Because the contents of the Type A granules do not
maintain their integrity at exocytosis, the contents must
be fluid; it is therefore hypothesized that the components
continuous with the egg surface. Two accessory cells (AC1 and AC2), the plasma membranes of which are
indicated by arrows, separate the more mature oocyte from a less mature one (O2). which has a thin primary
shell (TS2) but has not begun exocytosis of Type A granules (A2). Bar = 1.0 ^m. (B) The cortex of the egg
after shell formation. Clear peripheral vesicles (PV) occur at the surface of the egg (E) after the homogeneous
layer (HL) and fibrillar layer (FL) of the shell have formed around it. The vesicles communicate with
extracellular space. Asterisks mark fragments of cortical cytoplasm that may have been cut off from the
cytoplasm of the egg. Bar = 0.5 nm. (C) The mature shell of a prelaid egg. The egg (E) is surrounded by its
accessory cell (AC) and the thin primary shell (TS). Beneath the primary shell the homogeneous layer (HL)
and fibrillar layer (FL) of the mature shell encompass the egg. A homogeneous stripe (HS) can be seen
between the primary shell and the homogeneous layer. Although the plasmalemma of the egg is not clearly
seen here, adjacent micrographs show that it lies just interior to the fibrillar layer of the shell. Bar = 1.0 ^m.
(D) A representative section of the ooplasm of the egg after shell formation but before egg-laying. Type
B granules (B) and lipid droplets (L) are present, but Type A granules cannot be detected. Compare
Figure 5D with Figure IE, an area of similar size at a stage before formation of the mature shell. Bar
= 1.0 Mm.
f'-^-m!m?%®iB£.
x V-£«ssi&. ll«
6C
Figure 6. (A) Photomicrograph of two cleaving embryos surrounded by a still flexible, clear eggshell
(arrow). Bar = 10.0 ^m. (B) A magnified region of the eggshell of a laid embryo. Arrows indicate the layer
of the laid shell that probably corresponds to the primary shell, homogeneous stripe, and homogeneous
layer; these are no longer discrete. The fibrillar layer (FL) can still be discerned. The thickness of the shell
62
SHELL DEPOSITION IN AN ACOEL
63
flow to fill evenly the space between the primary shell and
the oolemma before hardening. The shell surrounding the
laid embryo shows little variation in thickness.
Because Type A granules are released by exocytosis and
profiles in which the membrane of the Type A granule is
continuous with the plasmalemma are common, it is likely
that the plasma membrane of the egg following shell for-
mation is a mosaic of the original oocyte membrane and
the membrane of the Type A granules. The occasional
fusion of Type A granules with one another rather than
with the plasmalemma could explain what appear to be
membrane-bounded fragments of cytoplasm that can
sometimes be found between the plasma membrane of
the egg and the newly synthesized shell. Because serial
sections were not taken, however, it is possible that the
apparent fragments are connected to the ooplasm in some
plane. The membranes of the clear peripheral vesicles as-
sociated with the plasmalemma soon after exocytosis of
the Type A granules are hypothesized to be the mem-
branes of the empty Type A granules.
The Type B granules were not observed to participate
in eggshell formation. Type B granules are hypothesized
to be the yolk granules, as yolk granules occur in the oo-
cytes of all archoophoran platyhelminths.
The observed increase in volume densities of the Type
B granules and the lipid droplets in the laid, cleaving em-
bryo when compared to the unlaid egg was initially per-
plexing. There is no indication of new synthesis of
granules. A decrease in volume of the embryo following
egg laying was hypothesized. Morphometric measure-
ments of unlaid eggs and laid, developing embryos of
C. pulchra demonstrated that the volume of the laid
embryo is significantly less than that of the unlaid egg.
This change in volume, which may well be a fixation
artifact, probably accounts for the apparent two-fold
increase in volume densities of Type B granules and
lipid droplets.
To our knowledge, the only turbellarians for which the
origin of eggshells has been examined ultrastructurally
are the rhabdocoels Microdalyellia fairchildi (Bunke,
1982) and Mesostoma ehrenbergii (Domenici and Gre-
migni, 1977) and the polyclads Pseudostylochus sp. and
Planocera multitentaculata (Ishida and Teshirogi, 1986;
Ishida, 1989). In the rhabodocoel Microdalyellia fairchildi,
some regions of the uterine epithelium release a vesicular
secretion against which EFGs from the yolk cells are se-
creted, although the exact role of this secretion in eggshell
formation is not clear (Bunke, 1982). Ishida and Teshirogi
(1986) describe a dual origin for the eggshells of the
polyclads in which an eggshell envelope is synthesized by
the shell glands in the female reproductive system, and
the remainder of the shell is formed following release of
the EFGs from the oocyte. Ishida (1989) has demonstrated
experimentally that the envelope is required for formation
of the eggshell. In C. pulchra the accessory cells to the
oocyte form a thin shell or envelope against which the
contents of the Type A granules are released. Thus it ap-
pears that an exogenous layer may be required to delineate
the parts of the shell formed from EFGs. The origin of
the external layer varies among the three groups of tur-
bellarians studied to date. Within the polyclads, however,
the origin of the external envelope is constant in the two
species examined. Only with additional studies will it be
possible to know if the origin of the external layer is a
useful taxonomic character.
EFGs in acoelonwrphans
The EFGs of C. pulchra are similar to the EFGs of
other turbellarians in three ways: (1) they contribute to
the formation of the eggshell, (2) they have a complex
morphology, and (3) the synthesis of EFGs begins prior
to the synthesis of yolk granules, as in other turbellarians
in which yolk granules are a product of the Golgi. The
EFGs of all acoelomorphans (Acoela and Nemertoder-
matida) studied to date differ from the EFGs of other
turbellarians in that the former lack polyphenols (see
Gremigni, 1 988; Smith et a!.. 1988; Falleni and Gremigni,
1990). Another apparent difference is that the EFGs found
in oocytes of acoelomorphans are smaller than those in
oocytes and vitellocytes of other turbellarians. EFGs of
most turbellarians are 1-2 ^m in diameter, whereas the
average diameter of the EFGs of C. pulchra is only one-
quarter to one-half as large, or 0.48 /*m. The EFGs of
other acoelomorphans appear like those of C. pulchra,
with diameters of approximately 0.5 /urn (see Gremigni,
1988; Smith et ai. 1988). Falleni and Gremigni (1990)
have recently suggested that the EFGs of "Convoluta
psammophyla" fuse as they migrate centripetally to form
granules 1-1.2 /um in diameter.
The species of acoels that have been examined for
eggshell formation belong to two different families, as-
suming that the "Convoluta psammophyla' examined by
Falleni and Gremigni (Gremigni, 1988; Falleni and Gre-
of the laid embryo is much more constant than that of the shell of the unlaid egg. Extra-embryonic space
(ES) separates the shell from the developing embryo. Bar = 1.0 ^m. (C) The shell (arrow) of the embryo
after the first two cleavages. Micromeres (Ml) and macromeres (MA) can be distinguished, and Type B
granules (B) and lipid droplets (L) occur in both types of blastomeres. The extensive extra-embryonic space
(ES) is shown. Bar = 10.0 jjm.
64
R. M. CHANDLER ET AL
migni, 1989, 1990) is, in fact, Paedomecynostomum
psammophilum (Family Mecynostomidae; see Beklem-
ischev, 1957; Dorjes, 1968). Differences in our results and
those of Falleni and Gremigni make a morphological
characterization of an EFG within the Acoela difficult at
this point. The electron-dense granules similar in size to
the Type A granules that form the eggshell in "Convoluta
psammophila" (Falleni and Gremigni, 1989, 1990) appear
to differ from those in C. piilchra in at least two ways.
First, the shell-forming granules of "C psammophila"
occur in clusters in the ooplasm, whereas they appear to
be randomly located in the oocytes of C. piilchra. Second,
in "C. psammophila" the granules marginate and fuse
with one another to form granules with a diameter 1-1.2
l/m. The granules of C. piilchra were not observed to fuse
with one another except occasionally at the time of exo-
cytosis. Certainly additional studies of eggshell formation
in acoelomorphans are required.
Phylogenetic implications
Smith el a/. (1986) suggested that the Turbellaria may
be polyphyletic with three distinct lineages: ( 1 ) the Ca-
tenulida, (2) the Nemertodermatida-Acoela [ = Acoelo-
morpha (Ehlers, 1984)], and (3) the Haplopharyngida-
Macrostomida-Polycladida-Neoophora and all higher
parasitic platyhelminths [ = Rhabditophora (Ehlers. 1984)].
The discovery that acoels' EFGs lack polyphenols
prompted examination of the composition of the EFGs
in the sister group to the acoels, the Nemertodermatida
(Thomas el ai. 1985). The oocytes of the nemertoder-
matid Nemertinoides elongatus were negative for poly-
phenols (Smith et a!.. 1988). The absence of polyphenols
from EFGs of the Acoelomorpha stands in sharp contrast
to the presence of polyphenols in EFGs in the Rhabdi-
tophora and provides further evidence supporting phyletic
distance between the Acoelomorpha and the Rhabdito-
phora. Yet to be discovered are characters that clearly
link the Acoelomorpha to the other groups (Smith et al,
1982). Studies of eggshell formation in the catenulids may
provide this link. There are several examples of homology
linking the Catenulida and the Rhabditophora, including
the ciliary rootlet system in their epidermal cells and the
origin of replacement cells for the epidermis in the pa-
renchyma (Ehlers, 1984; Smith el al.. 1986). Oocytes of
catenulids, which have an eggshell that arises from gran-
ules within the oocyte (Borkott, 1970), have never been
examined by electron microscopy. It will be interesting
to examine the morphology of catenulid EFGs and to
find out if they contain polyphenols. If the Catenulida do
have polyphenolic EFGs, this fact would further separate
the Catenulida and Rhaditophora from the Acoelomor-
pha. Although one should consider the possibility that the
lack of polyphenols in EFGs is derived, if the catenulids
have a non-polyphenolic EFGs with the same morphology
as the EFGs of C. piilchra, the morphology of the EFG
could provide a link between the Acoelomorpha and the
Catenulida. A non-polyphenolic EFG could then repre-
sent a plesiomorphy for the Platyhelminthes as suggested
by Falleni and Gremigni (1989, 1990).
Acknowledgments
The authors thank Dr. Lawrence S. Barden and Dr.
Larry Leamy for help with statistical analyses and Ms.
Sandra F. Zane for assistance in electron microscopy.
Laboratory space was kindly provided by the North Car-
olina Aquarium at Fort Fisher, NC. The work was sup-
ported in part by a Sigma Xi Grant-in-Aid (to RMC) and
in part by a UNCC Faculty Research Grant (to MBT).
Literature Cited
Beklemischev, V. N. 1957. Convoluta psammophila sp. nov. and the
tendency toward juvenile oligomerisation of cellular elements in
Turbellaria Acoela. 7>. Leningr. Ova. Estestvoispyt. 73: 5-13. [In
Russian].
Borkott, H. 1970. Geschlectliche Organisation, Fortpflanzungsverhal-
ten und Ursachen der sexuellen Vermehrung von Stenostomum
sthenum nov. spec. (Turbellaria, Catenulida). /. Morphol. Tiereftl:
183-262.
Boyer, B. C. 1972. Ultrastructural studies of differentiation in the oocyte
of the polyclad turbellanan. Prostheceraeus floridanus. J. Morplioi
136: 273-296.
Bunke, D. 1982. Ultrastruktur-untersuchungen zur Eischalenbildung
bei Microdalyellia fairchildi (Turbellaria). Zoomorphology 101: 61-
70.
Chandler, R. M., and M. B. Thomas. 1986. The ultrastructure of the
oocyte of Convoluta sp. J Elisha .Mitchell Sci. Soc. 102: 179-180
(abstr.).
Chandler, R. M., and M. B. Thomas. 1987. An Ultrastructural exam-
ination of late oogenesis and of the laid egg of Convoluta sp. Am.
Zoo/. 27: 121 A (abstr.).
Chandler, R. M., J. P. Smith, and M. B. Thomas. 1988. A morpho-
metric analysis of egg shell formation in Convoluta sp. J. Elislia
Mitchell Sci. Soc. 104: 190 (abstr.).
Coil, \V. H. 1991. Platyhelminthes: Cestoidea. Pp. 21 1-283 in Micro-
scopic Anatomy of Invertebrates, vol. 3: Platyhelminthes and Nem-
c'rtinea. F. W. Harrison and B. J. Bogitsh. eds. Wiley-Liss. New York.
Dorjes, J. 1968. Die Acoela (Turbellana) der Deutschen Nordseekiiste
und ein neues System der Ordnung. Z. Zoo/. Syst. Evolutionsforsch,
6: 56-452.
Domenici, L., and V. Gremigni. 1977. Fine structure and functional
role of the coverings of the eggs in Afesostoma ehrenbergii (Focke)
(Turbellaria, Neorhabdocoela). Zoomorphologie 88: 247-257.
Domenici, L., L. Galleni, and V. Gremigni. 1975. Electron microscopical
study of egg-shell globules in Notoplana alcinoi. J Submicrosc. Cytol.
7: 239-274.
Ehlers, L'. 1984. Das Phylogenetische System der Plathelminthes.
Gustav Fischer Verlag. Stuttgart. 3 1 7 pp.
Espinosa, P. L. 1986. Histochemical and Ultrastructural studies of oo-
genesis in the marine polyclad turhellarian Gnesioceros floridana.
M.S. Thesis. The University of North Carolina at Charlotte, Charlotte.
NC.
Falleni, A., and V . Gremigni. 1989. Egg covering formation in the acoel
Convolula psammophyla (Platyhelminthes, Turbellaria): an ultra-
SHELL DEPOSITION IN AN ACOEL
65
structural and cytochemical investigation. Ada. Embryo/. Morphol
Exper. N.S 10: 105-112.
Kalleni, A., and V. Gremigni. 1990. LHtrastructural study of oogenesis
in the acoel turbellarian Convoluta. Tissue & Cell 22: 301-310.
Fried, B., and M. A. Haseeb. 1991. Platyhelminthes: Aspidogastrea,
Monogenea, and Digenea. Pp. 141-209 in Microscopic Anatomy of
Invertebrates, vol. 3: Platyhelminthes and Nemertinea, F. W. Harrison
and B. J. Bogitsh, eds. Wiley-Liss, New York.
Gammon, C. 1979. Histochemical and ultrastructural studies of oo-
genesis in the marine polyelad turbellarian Stylochus zebra Verrill.
M.S. Thesis, Wake Forest University, Winston-Salem, NC.
Giesa, S. 1966. Die Embryonalentwicklung von Monocelis jitsca
Oersted (Turbellaria. Proseriata). Z Morphol. Oekol. TiereSl: 137-
230.
Gremigni, V. 1988. A comparative ultrastructural study of homocellular
and heterocellular female gonads in free-living Platyhelminthes-Tur-
bellaria. Fortschr. Zoo/. 36: 245-261.
Gremigni, V., A. Falleni, and P. Lucchesi. 1987. An ultrastructural
study of oogenesis in the turbellarian Macroslomum. Ada Embryol.
Morphol. Exper. N.S. 8: 257-262.
Hulings, N. C., and J. S. Gray. 1974. A manual for the study of mei-
ofauna. Smithson. Contr. Zoo/. 78: 1-83.
Ishida, S. T. 1989. Further studies on the shell-forming granules and
eggshell formation in polyclads (Turbellaria, Platyhelminthes). Sci.
Rep. Hirosaki Univ. 36: 55-72.
Ishida, S. T., and \V. Teshirogi. 1986. Eggshell formation in polyclads
(Turbellaria). Hydrobiologia 132: 127-135.
Ishida, S. T., T. Gotoh, and \V. Teshirogi. 1981. Oogenesis and egg-
shell formation in polyclads. Rep. Fukaura Mar. Bio/. Lab., Hirosaki
Univ. 9: 32-48. [translated from Japanese].
Kucera, F. P. 1987. An ultrastructural and cytochemical analysis of
eggshell-forming granules in Macroslomum beaufortensis and Ma-
croslomum hystricimim marinum (Turbellana, Macrostomida). M.S.
Thesis, The University of North Carolina, Charlotte, NC.
McDowell, E. M., and B. F. Trump. 1976. Histological fixative suitable
for diagnostic light and electron microscopy. Arch. Pathol. Lab. Med.
100: 405.
Reynolds, E. S. 1963. The use of lead citrate at high pH as an electron
opaque stain in electron microscopy. J. Cell Biol. 17: 208.
Rieger, R. M., S. Tyler, J. P. S. Smith, and G. E. Rieger. 1991 . Platy-
helminthes: Turbellana. Pp. 7-140 in Microscopic Anatomy of In-
vertebrates, vol. 3: Platyhelminthes and Nemertinea, F. W. Harrison
and B. J. Bogitsh. eds. Wiley-Liss, New York.
Smith, J. P. S., and L. Bush. 1991. Convolula pulchra n.sp. (Turbellaria:
Acoela) from the East Coast of North America. Trans. Am. Alicrosc.
Soc. 110: 12-26.
Smith, J. P. S., S. Tyler, M. B. Thomas, and R. M. Rieger. 1982. The
morphology of turbellarian rhabdites: phylogenetic implications.
Trans. Am. Microsc. Soc. 101(3): 209-228.
Smith, J. P. S., S. Tyler, and R. M. Rieger. 1986. Is the Turbellaria
polyphyletic? Hydrobiologia 132: 13-21.
Smith, J. P. S., M. B. Thomas, R. M. Chandler, and S. F. Zane. 1988.
Granular inclusions in the oocytes of Convoluta sp., Nemertoderma
sp., and Nemertinoides e/onga/its (Turbellaria, Acoelomorpha).
Fortschr. Zoo/. 36: 263-269.
Spurr, A. R. 1969. A low-viscosity epoxy resin embedding medium
for electron microscopy. J. Ultrastruct. Res. 26: 31-43.
Thomas, M. B., J. P. S. Smith, R. M. Chandler, and A. Barker. 1985.
Eggshell granules in some primitive Turbellaria: more evidence for
polyphyly? Am. Zoo/. 25: 91 A.
Watson, M. L. 1958. Staining of tissue sections for electron microscopy
with heavy metals. / Biophys. Biochem. Cytol. 4: 475-479.
Weibel, E. R. 1979. Slereological Methods, vol. I. Academic Press.
Inc., London. 415 pp.
Reference: Biol. Bull. 182: 66-76. (February, 1992)
Isolation of Neurons of a Nudibranch Veliger
FU-SHIANG CHIA, RON KOSS, SHAUNA STEVENS, AND JEFF I. GOLDBERG
Department of Zoology, University of Alberta, Edmonton, Alberta, Canada T6G 2E9
Abstract. A technique was developed to dissociate and
culture identified sensory neurons and interneurons from
the anterolateral propodial ganglia of metamorphically
competent veligers of the nudibranch, Onchidoris bilain-
ellata. Receptor cells have been represented as receiving
an environmental cue that initiates the settlement re-
sponse. The ganglionic cells, along with other cell-types
from the propodial region housing the ganglia, were ex-
cised with a large-bore micropipette, and dissociated by
mild trypsin incubation and trituration. Cells and tissues
were plated in poly-L-lysine-coated plastic culture dishes
containing modified Aplyxia medium and survived for up
to four days. The different cell-types possess diagnostic
features, so they can be recognized under culture condi-
tions. Sensory cells were bipolar in profile, with one end
of the cell being thickened, representing the receptor ap-
paratus. Interneurons are unipolar or bipolar in shape,
and bear thin neurites. Other cell-types, including myo-
cytes, ciliated epidermal cells, nonciliated epidermal cells,
and gland cells were identified. Identifications of living
cells were corroborated through electron microscopical
analysis.
Introduction
The propodium of the advanced veliger larva of the
nudibranch Onchidoris bilainellala contains a unique set
of morphologically identifiable structures called the an-
terolateral ganglia (Chia and Koss, 1989). These structures
are thought to be involved in perception of the environ-
mental cues that induce settlement and metamorphosis
(Chia and Koss, 1988). Recently, it has been shown that
sensory receptor cells within the anterolateral ganglia re-
spond to a known settlement cue (barnacle-conditioned
seawater) by producing slow, low-amplitude depolariza-
tions that can be detected by the application of conven-
Received 10 September 1990: accepted 25 November 1991.
tional intracellular recording methods (Arkett el ai, 1989).
However, the activity of the sensory cells is variable in
terms of duration and amplitude. Such variability may
be due to the changeable concentration of crude stimulus
being administered, the developmental status of the re-
ceptors, or the degree of damage caused by the electrode
upon entry into the cell. Also unclear is the depth within
the ganglion that the electrode was placed, and whether
the epidermal tissues overlying these subepidermal struc-
tures altered the response of receptor cells.
In this paper we report a technique for excising, dis-
sociating, and culturing the lateral propodial region of the
veliger foot of O. bilamel/ata. This work was undertaken
with a view to studying the responses of the isolated cells
to settlement or metamorphic cues, thereby overcoming
many of the limitations presented by the intact prepara-
tion.
Materials and Methods
Egg ribbons from Onchidoris bilainellala (Linnaeus,
1767) were collected at Bamfield Marine Station, Barn-
field, British Columbia, Canada, and transported to the
University of Alberta. Veliger larvae were then raised ac-
cording to the method of Chia and Koss ( 1988). The ve-
ligers used in these experiments were from the same co-
hort, and random samples from the cohort were checked
for metamorphic competency according to Chia and Koss
(1988).
Veligers were pipetted into a Sylgard (Dow Corning)-
lined dish containing 2.5 ml of a high Mg++, low Ca++
seawater mixture (12-15°C) consisting of natural sea-
water, isotonic (0.33 M) MgCl-., and Co++-seawater in a
ratio of 2:1:4.5 (v/v/v). Co++-seawater consisted of 430
mM NaCl, 10 mA/CoCl:, 10 mM KC1, 30 mM MgCl2,
20 mM MgSOj, 10 mA/TES pH 7.8 (Arkett et ai, 1987,
1989). Veligers were tethered with a cactus spine, which
was inserted through the base of the velum. A second
66
CULTURE OF LARVAL CELLS
67
spine was placed through the tip of the foot, and a third
spine was used to orient the veliger such that one side
of the propodium and receptor field faced upward (Figs.
1,2).
Larvae were subsequently incubated for 5 min in 0.2%
trypsin (Sigma) in an Aplysia denned medium [mADM:
50% Liebowitz L- 1 5 (Gibco special order); 0.26 M NaCl;
9.7 mAl CaCV. 4.6 mAf KC1; 26 mA/ MgSO4; 26 mM
MgCl:; 2 mAfNaHCO3; 33 mA/ Dextrose; 10 mA/Hepes;
0.015% L-glutamine; 50 ^g per ml gentamicin] modified
from Schacher and Proshansky (1983). This solution was
eventually replaced with mADM.
The anterolateral ganglion, which is visible as an oblong
cellular mass, was located using a dissecting microscope,
and a silicon-coated (Sigmacoat) micropipette, with a bore
diameter of 20-30 jum, was placed directly on it. A mi-
cromanipulator was used to place the micropipette. The
tissue was excised by first applying mechanical force with
the micropipette to this region, and then alternating neg-
ative and positive pressure within the micropipette
through a microsuction device (Canlab). The excised tis-
sues, including the portions of the anterolateral ganglion
and propodial epidermis, were triturated in the micro-
pipette. The dissociated cells were then plated on to high
molecular weight poly-L-lysine (Sigma) coated plastic 35
mm tissue culture plates (Falcon) in cold mADM. The
cultures were maintained at 4°C for 24 to 72 h.
Cell cultures were viewed and photographed live with
a Nikon TMD inverted photomicroscope equipped with
phase-contrast optics. Cultures to be fixed for transmission
electron microscopy (TEM) were stained with Richard-
son's stain (Richardson el a/., 1960) to locate the cells in
the culture dish. For TEM, cells were fixed for 1 h in 2.5%
glutaraldehyde in 0.2 M phosphate buffer (pH 7.6), fol-
lowed by a 1 h post-fixation in 2% OsO4 in 1.25% sodium
bicarbonate (pH 7.2, Wood and Luft, 1965). They were
then dehydrated through an ethanol series, and directly
embedded in Medcast (Ted Pella Inc.). After a polymer-
ization period of 72 h, the Medcast was peeled from the
culture dish, mounted on Medcast blanks with Krazy
Glue, and sectioned with a diamond knife. Sections were
stained with uranyl acetate and lead citrate (10 min each).
Sections were examined with a Philips E. M. 201 electron
microscope.
For scanning electron microscopy (SEM), larvae were
relaxed in the Co++-seawater mixture described above,
and processed according to the technique of McEuen
(1985). Specimens were then examined with a Cambridge
Stereoscan 250 SEM.
Results
The foot of the advanced veliger of O. bilamellala is a
large structure consisting of a metapodium and propo-
dium (Figs. 1, 2). In veligers that have been immobilized
in the high Mg++, low Ca++ seawater, the location of each
anterolateral ganglion is evident as an oblong, cilia-free
region on the foremost sides of the propodium (Fig. 2).
The sensory fields overlying the anterolateral ganglia are
composed of a mosaic of cell-types including the dendrites
of sensory cells, epidermal cells, multicellular metapodial
glands, and muscle cells (Figs. 3-5). Some untargeted tis-
sues and cells were inevitably excised along with the target
tissues, and will be characterized because they are present
in cultures of dissociated cells.
The pair of anterolateral propodial ganglia are located
below the level of the epidermis (Figs. 3, 5). They are
oblong structures, consisting of an outer cortex of cell
bodies and an inner neuropil of fibers (Chia and Koss,
1989). The cell-types constituting the cortex include sen-
sory cells, neurons, and sheath cells; neurons occupy the
inner layer adjacent to the neuropil, while sensory cells
are more numerous and distributed around the outer, lat-
eral perimeter of each ganglion. Bundles of dendritic pro-
cesses originating from sensory cells traverse the epidermis
to form the external sensory fields mentioned above (Figs.
1, 2, 5). Sheath cells encapsulate the ganglia. The antero-
lateral ganglia are connected to the central nervous system,
i.e.. cerebral ganglia, by short commissures.
After excision of a ganglion, the dissociated cells settled
within 24 h of plating. Cells settled in clusters or as in-
dividuals and attached to the substratum. Several cell-
types were identifiable at the level of the light microscope
on the basis of their size and morphology. These included:
epidermal cells (both ciliated and nonciliated). gland cells,
muscle cells, neurons, and sensory neurons (Figs. 7, 10,
13, 16, 19, 22, 25, 29). The appearance of these cells was
constant in several separate dissociations and was com-
parable to those identified by in situ study. Most impor-
tantly, the fine structure of the different cell-types in situ
was conserved under in vitro conditions, thus corrobo-
rating the identification of cell-types according to light
microscopy.
Epidermal cells
In its natural state, the epidermis covering the antero-
lateral ganglion is composed primarily of multiciliated
cells and cells with long branched microvilli projecting
from their apices (Figs. 5,6,9). Both cell-types are cuboidal
in profile and attach to the basal lamina through a hemi-
desmosome complex involving numerous microfilaments.
Those cells that bear microvilli contain numerous mito-
chondria within the apical portion of the cell; electron-
dense granules of about 0.2 ^m in diameter are found
immediately below these organelles (Fig. 6). Multiciliated
cells also possess numerous distal mitochondria, which
**»,
^
,
SB ••••
-»
68
CULTURE OF LARVAL CELLS
69
are distributed among the ciliary rootlets. Nuclei occupy
the basal regions of both cell-types. However, the cyto-
plasm and the nuclei of microvillar cells stain less densely
than ciliated cells, while the latter possess unique vacuoles
with electron-translucent contents.
Immediately after dissociation of a ganglion, cultures
are dominated by ciliated and non-ciliated epidermal cells,
many of which fail to attach to the poly-L-lysine substra-
tum. Ciliated and nonciliated epidermal cells were readily
recognizable at the light microscopy level because both
types of cell were round (Figs. 7, 10). In ciliated cells, the
cilia were readily identifiable (Fig. 10). After trituration,
both of these types of cell generally remained aggregated.
The structural characteristics of the in situ and //; vitro
ciliated and nonciliated epidermal cells are similar; those
cells found in culture are round to cuboidal and contain
microfilaments, numerous mitochondria, and microvilli
or cilia (Figs. 5, 6, 8, 9, 1 1 ). The substructure of the cilium
includes a ciliary rootlet. Within a day, cilia appear to be
reabsorbed into the cell. In contrast, nonciliated epidermal
cells contain characteristic electron-opaque granules of
about 0.5 jim in diameter, and the nucleus stains lightly.
Muscle cells
Muscle cells are numerous throughout the propodium
and lie adjacent or attach to the basal lamina. They are
filamentous in profile and their contents are dominated
by myofilaments (Figs. 3, 4, 12). Mitochondria are scat-
tered around a centrally positioned elliptical nucleus,
which is displaced away from the main stream of myo-
filaments (Fig. 12). Individual muscle cells found under
culture conditions are identical in shape and structure.
They are large and filamentous in appearance and their
nucleus protrudes from one side of the cell (Figs. 13, 14).
The cytoplasm contains bundles of myofilaments which
extend the length of the cell, and mitochondria which are
scattered around the nucleus (Fig. 14).
Secretory cells
The metapodial glands of intact ganglia are flask-shaped
structures and are located below the level of the epidermis
(Figs. 3-5). They contain two types of gland cells (Figs.
4, 15, 18). The first, or type A secretory cell, contains
inclusions that are composed of flocculent material; free
ribosomes and rough endoplasmic reticulum are domi-
nant organelles and surround an oblong nucleus (Fig. 15).
The second, or type B secretory cell has a densely staining
nucleus, and its cytoplasm is filled with small electron-
dense granules averaging 0.1 ^m in diameter (Fig. 18).
Isolated metapodial gland secretory cells had the same
overall morphology; they were large and bulbous, with
the nucleus occupying the rounded end of the cell (Figs.
15, 17, 18, 20). In certain instances, large secretory vac-
uoles in type A cells gave the interior of the cell a con-
voluted appearance when observed by light microscopy.
Isolated type A secretory cells contained the vacuoles and
extensive rough endoplasmic reticulum observed in situ
(Fig. 17). Cultured type B cells contained the electron-
dense granules present in situ and could often be distin-
guished from type A cells by their smaller cytoplasmic
protrusions (Fig. 20).
Ganglionic cells
Anterolateral ganglia are composed of sheath cells,
neurons (Figs. 21, 24), and sensory receptor cells (Figs.
27, 28) whose configuration, distribution, and fine struc-
ture have been described in detail by Chia and Koss
(1989). Sheath cells stain densely and encapsulate the an-
terolateral ganglia through long, slender processes. How-
Figure I. Scanning electron micrograph (SEM) of an advanced veliger showing location of one of the
pair of settlement receptor fields (arrow), propodium (P), metapodium (M). shell (H), velum (V), and oper-
culum (O). Bar = 100 pm.
Figure 2. Higher magnification SEM showing a receptor field (arrows) and the foot with propodium (P)
and metapodium (M). Bar = 20 ^m.
Figure 3. Section ( 1 fim thickness) through the propodium identifying the position of the anterolateral
ganglia (A), covering epidermis (EP), operculum (O), and opercular muscle (OM). The multilobular propodial
glands (PG) and the smaller metapodial glands (MG) are also shown. Bar = 10 jum.
Figure 4. Transmission electron micrograph of section through a ciliated region of the propodium showing
the epidermis composed of ciliated cells (CO, overlying metapodial glands ( MG). and muscle cells and fibers
(ME). Note that the secretory cells within the metapodial glands are of two types; one type has large electron-
translucent inclusions (G) and the other type has darkly staining nuclei and small electron-dense granules
(D). Bar = 5 //m.
Figure 5. Cross section through the lateral propodial region showing that the anterolateral ganglion (A)
is composed of sensory cells (S) with radiating ciliated dendritic endings (SD). interneurons (N). and sheath
cells (E). The epidermis (EP) contains ciliated cells (CC), microvillar cells (MA) with elongate microvilli,
and the ramifications of sensory cell terminals. Also note the basal lamina (BL), muscle cell (ME) and
metapodial gland cell processes (arrow). Bar = 2.5 /jm.
•
X
> -i
,
- . •- - • J, -t-
L1
Figure 6. Section through an in situ microvillus cell located in the epidermis overlying the anterolateral
ganglion. Note the nucleus (NU), electron-dense granules (G) and microvilli (arrow). Bar = 1 jim.
Figure 7. Phase-contrast photomicrograph of a cultured microvillus cell isolated from the lateral propodial
region of a competent Onchidoris veliger. Bar = 10 nm.
70
CULTURE OF LARVAL CELLS
71
ever, isolated sheath cells were never found in cultures of
dissociated cells, although they were rarely observed at-
tached to undissociated clusters of neurons and sensory
cells.
Neurons are located deep within an intact ganglion.
Their cell bodies contain a relatively clear cytoplasm, a
lightly staining nucleus, free ribosomes, and a small num-
ber of vesicles (Figs. 2 1 . 24). They possess one or two thin
neurites, which, depending upon the location of that neu-
ron within the ganglion, projected from the cell body (Fig.
24) into the neuropil or the commissure that connects to
the cerebral ganglion. In situ neurites are visualized as
thin processes that contain a few vesicles, microtubules,
and the occasional mitochondrion (Fig. 24); neurons do
not communicate with the external environment and do
not possess cilia.
Neurons plated from the anterolateral ganglion were
identified by the presence of long thin, sometimes
branched processes that extended from a teardrop-shaped
unipolar or bipolar cell body (Figs. 22, 25, 29). They were
usually found in loose aggregates along with sensory cells,
with unipolar or flask-shaped interneurons being the most
numerous (Fig. 25). The soma of interneurons ranged
from 3-5 /urn in diameter, which is the size range of the
neuronal soma identified in situ. In sectional profile, both
in situ and isolated neurons possessed long slender neurites
that emerge from a bulbous base containing the nucleus
(Figs. 21, 24, 26). The nucleus was spherical to cuboidal
in form and embedded, along with a few mitochondria,
in a finely granular cytoplasm (Figs. 22-25). Neurites
contained microtubules and a small number of vesicles
ranging from 50 to 80 nm in diameter (Figs. 24, 26). All
the above features are identical to those of neurons in situ
in the anterolateral ganglia of the veliger (Chia and Koss,
1989). No synaptic profiles were observed within the gan-
glion, although synapses were observed in the connectives
that connect to the cerebral ganglion. Similarly, synapses
were never observed in dissociated cells, and it was not
possible to classify them as motoneurons or interneurons.
The cell bodies of sensory cells in situ are located within
the ganglion along the lateral margin that borders the epi-
dermis (Figs. 5, 27). They are spindle-shaped, and consist
of a lightly staining nucleus, free ribosomes, microtubules,
and numerous vesicles, all of which are embedded in a
relatively clear cytoplasm (Fig. 27). Axonal processes,
which project to the neuropil, contain microtubules, ves-
icles, and the occasional mitochondrion. The character-
istic features of sensory neurons include: ( 1 ) dendritic
processes that are thicker than axons, (2) dendrites that
extend from the cell body through the epidermis to ter-
minate externally as a single cilium. The cilium was ob-
served to possess a basal body but no rootlet, (3) large
(0.5 jum) electron-dense granules, and (4) numerous ves-
icles in the cell body. Their dendrites also contain mito-
chondria and microtubules. The overall appearance (Chia
and Koss, 1989) of these cells is retained following dis-
sociation and plating in the culture dishes (Figs. 27, 29,
30). Isolated sensory neurons were recognizable at the
level of light microscopy by their characteristic spindle to
cigar shape: one of the two processes that radiated out
from the soma was thicker, and often shorter, and pre-
sumably represented the dendrite (Fig. 29). The nucleus
occupied a central location within the cell (Fig. 29). At
the fine structural level, the integrity of the sensory cells
also remained unchanged. These cells are characterized
by a relatively clear, lightly staining cytoplasm that con-
tains free ribosomes and an elliptical to spherical nucleus
(Figs. 27, 28, 30). The nucleus is relatively lightly staining,
and the axon and dendritic processes can be recognized.
The dendrite possesses a single cilium and mitochondria,
is generally wider than the axon, and characteristically
contains vesicles and electron-dense granules averaging
0.5 Mm in diameter (Figs. 27, 28, 30). The cilium lacks a
ciliary rootlet as compared to cilia of general ciliated epi-
dermal cells which have rootlets (Fig. 30). However, the
cilium could not always be found in all cells suspected of
being sensory cells, possibly because this organelle may
sometimes have been reabsorbed or truncated after the
dissociation. The axons contain microtubules and vesicles
and are similar in appearance to the neurites of interneu-
rons described above.
The diagnostic features of the different cell-types found
in intact ganglia and in cultures of dissociated ganglia are
summarized in Table I.
Figure 8. Section through an isolated microvillus cell, under in vitro conditions, showing microvilli
(arrows) electron-dense granules (G) and nucleus (NU). Bar = 1 jim.
Figure 9. In situ ciliated epidermal cell showing multiple cilia (C) and a densely staining nucleus (NU).
Bar = 1 ftm.
Phase-contrast photomicrograph of a cultured ciliated epidermal cell showing multiple cilia
Figure 10.
(C). Bar = 10
Figure 11.
Figure 12.
Figure 13.
Figure 14.
In vitro ciliated cell showing multiple cilia (C) and nucleus (NU). Bar = 1 ^m.
In situ muscle cell showing nucleus (NU) and myonlaments (MF). Bar = 1 ^m.
Phase-contrast photomicrograph of an isolated, cultured muscle cell. Bar = 5 Mil-
In vitro muscle cell showing myonlaments (MF) and nucleus (NU). Bar = 1 nm.
Figure 15. Type A gland cells of an in situ metapodial gland showing secretory granules (SG) filled with
flocculent material, rough endoplasmic reticulum (R). and a nucleus. Bar = 1 ^m. Inset: higher magnification
of secretory granules.
Figure 16. Phase-contrast photomicrograph of an isolated, cultured gland cell. Bar = 10 ^m.
CULTURE OF LARVAL CELLS
73
Table I
Identifying characteristics of cell-types from the anterolateral ganglionic region <>/ Onchidoris bilamellata
Cell-type
Form Staining Size
(*LM and TEM profile) properties (largest dimension)
Characteristic inclusions and organelles
Metapodial gland cell A
Metapodial gland cell B
Muscle cell
Epidermal ciliated cell
Epidermal microvillus cell
Sensory neuron
Neuron
club
flask
thread
cube
cube
spindle; one end thicker
and often shorter
oval bipolar or unipolar;
long slender neurites
lightly granular lO-lS^m large granules containing flocculent
material; rough endoplasmic reticulum
densely granular 1 5 p.m electron-dense granules (0. 1 nm diameter)
densely granular 20-50 pm myofilaments
densely granular 5-10 ^m multiple cilia; numerous mitochondria
lightly granular 5-10 ^m microvilli; opaque granules (0.2 jim
diameter)
lightly granular 7-10 ^m single cilium; microtubules; clustered
vesicles (0.05-0.08 ^m diameter);
opaque granules 0.5 ^m diameter
lightly granular 10-15 ^m microtubules; few vesicles (0.05-0.08 ^m
diameter)
* LM = light microscopy
Listed features are shared
TEM = transmission electron microscopy.
between identical cell-types found under in situ and in vitro conditions.
Discussion
Assays that measure whole-organism responses are
useful in identifying substances that induce larval settle-
ment and metamorphosis. However, they provide little
information about the cellular mechanisms that they ac-
tivate, or the location (i.e., the structural site) at which
they are detected (reviewed by Pawlik, 1990). Therefore,
techniques must be developed which enable the study of
the precise actions of settlement and metamorphic in-
ducers.
There is considerable evidence suggesting that the in-
duction of settlement and metamorphosis of larval mol-
lusks is mediated by the nervous system (Hadfield, 1978);
at the primary level, the process of perceiving natural or
artificial inductive substances is ascribed to an external
epidermal sensory cell (Baloun and Morse, 1984; Burke.
1983; Morse and Baxter, 1989; Morse, 1990; Trapido-
Rosenthal and Morse. 1986; Yool el al, 1986). Receptor
cells for these responses in different species of veligers
have been localized to the cephalic sensory organ (Bonar,
1978; Morse et al., 1980; Chia and Koss, 1984), rhino-
phores (Chia and Koss, 1982), and the anterior portion
of the foot or propodium (Chia and Koss, 1989). Although
the receptive capacities of most of these organs have been
inferred from morphological characteristics and relation-
ships, there is little evidence to suggest that all these struc-
tures are predisposed to perceive settlement or metamor-
phic cues. To date, the Onchidoris larval foot, or specif-
ically the anterolateral propodial ganglia, represents the
only system where morphologically identified chemosen-
sory receptor cells have been shown electrophysiologically
to respond to a known settlement cue (Arkett et al.. 1989).
In this study we have developed a method for disso-
ciating, isolating, and maintaining the cells that constitute
the anterolateral portion of the Onchidoris foot, including
the sensory neurons and interneurons of the anterolateral
ganglia. This technique produces cultures of individual
cells, permitting future studies of the cellular mechanisms
involved in the detection and transduction of settlement
cues; previous studies have been restricted by in situ prep-
arations.
This in vitro system is unique among preparations for
studying veliger settlement because it can be used to study
Figure 17. In vitro gland cell showing secretory granules (SG), rough endoplasmic reticulum (R), and
nucleus (N). Bar = 1 ^m.
Figure 18. Type B secretory cell showing a densely staining nucleus (N) and small electron-dense granules
(G). Bar = 1 ^m.
Figure 19. Phase-contrast photomicrograph of an isolated, cultured type B gland cell. Bar = 10 Aim.
In vitro type B gland cell showing small electron-dense granules and a densely staining nucleus.
Figure 20.
Bar = 1 nm.
Figure 21.
Figure 22.
Interneuron located in anterolateral ganglion. Bar = 1 /im.
Phase-contrast photomicrograph of a interneuron isolated from the anterolateral ganglion.
Note the bipolar shape with neurites (NU) radiating out from the cell body. Bar = 5 ^m.
Figure 23. Section through an interneuron under in vitro conditions. Bar = 1 nm.
Figure 24. Neurite (NU) of interneuron in an anterolateral ganglion showing vesicles and microtubules.
Bar = 1 Mm.
Figure 25. Phase-contrast photomicrograph of an isolated interneuron showing a unipolar shape with
a single neurite (NU) radiating from the cell body. Bar = 5 /jm.
Figure 26. Section through an isolated and cultured interneuron showing a neunte (NU) extending out
from the cell body. Inset: neurite with vesicles (arrows) and microtubules. Bars = 0.5 urn.
74
CULTURE OF LARVAL CELLS
75
the immunocytological properties of different cell types,
their passive and active electrical properties, and their re-
sponses to bioactive substances such as neurotransmitters
and neuromodulators. A similar approach has been suc-
cessfully employed to study the same parameters in neu-
rons dissociated from embryos of the pulmonate, Heli-
soma trivolvis (Goldberg ct a/., 1988: Goldberg and Price,
1 991; Goldberg el al, 1991).
For the most part, the isolated cells retained their basic
in situ appearance, and could thus be classified according
to cell-type by light microscopy. Such classifications were
usually confirmed by electron microscopy. There were
instances, however, where the appearances of different
cell-types overlapped sufficiently to preclude their classi-
fication. For instance, the apical ends of many putative
sensory cells apparently became rounded following iso-
lation, making them indistinguishable from gland cells.
Nevertheless, future studies that combine electrophysio-
logical techniques with those employed here, will ensure
accurate diagnosis of each cell-type.
Arkett el al. (1989) have demonstrated that settlement
receptor cells in Onchidoris depolarize in response to a
known settlement cue. It is our intention to use the culture
system developed in the present study to cultivate the
sensory cells and neurons of the anterolateral ganglion
for the purpose of voltage- and current-clamping exper-
iments. The electrophysiological effects of settlement in-
ducing-ligands and neurotransmitters can then be directly
and precisely monitored on single identified cells. The
possible roles of second messengers in mediating settle-
ment and metamorphic responses can then be accurately
elucidated.
Acknowledgments
The collection of egg masses, by Dr. A. Martel and M.
Sewell, was greatly appreciated. We also wish to thank
Dr. D. A. Craig and G. D. Braybrooke for providing
Scanning Electron Microscope Facilities, and Dr. S. K.
Malhotra for providing Transmission Electron Micro-
scope Facilities. F.-S. C. was supported by NSERC grant
#6083 and J.I.G. was supported by NSERC grant #U0553
and the Alberta Heritage Foundation for Medical Re-
search.
Literature Cited
Arkett, S. A., G. O. Mackie, and C. L. Singla. 1987. Neuronal control
of ciliary locomotion in a gastropod veliger (Calliostoma). Biol. Bull.
173:513-526.
Arkett, S. A., F.-S. Chia, J. I. Goldberg, and R. Koss. 1989. Identified
settlement receptor cells in a nudibranch veliger respond to specific
cue. Biol. Bull. 176: 155-160.
Baloun, A. J., and D. E. Morse. 1984. Ionic control of metamorphosis
in larval Haliotis rufescens (Gastropoda). Biol. Bull. 167: 124-138.
Bonar, D. B. 1978. LUtrastructure of a cephalic sensory organ in larvae
of the gastropod Pliesli/la sibogae (Aoelidacea, Nudibranchia). Tissue
Cell 10: 153-165.
Burke, R. B. 1983. The induction of metamorphosis of marine inver-
tebrate larvae: stimulus and response. Can. J. Zool. 16: 1701-1719.
Chia, F.-S., and R. Koss. 1982. Fine structure of the larval rhinophores
of the nudibranch. Rustanga pu/chra. with emphasis on the sensory
receptor cells. Cell Tiss. Res. 225: 235-248.
Chia, F.-S., and R. Koss. 1984. Fine structure of the cephalic sensory
organ in the larva of the nudibranch Rostanga piilcliru (Mollusca,
Opisthobranchia. Nudibranchia). Zoomorphology 104: 131-139.
Chia, F.-S., and R. Koss. 1988. Induction of settlement and meta-
morphosis of the veliger larva of the nudibranch, Onchidoris bila-
nwllata. Int. J. Invert. Reprod. 14: 53-70.
Chia, F.-S., and R. Koss. 1989. The fine structure of the newly dis-
covered propodial ganglia of the veliger larva of the nudibranch,
Onchidoris bilamellata. Cell Tiss. Res. 256: 1 7-26.
Goldberg, J. I., D. P. McCobb, P. B. Guthrie, R. A. Lawton, R. E. Lee,
and S. B. Kater. 1988. Characterization of cultured embryonic
neurons from the snail Helisoma. Pp. 85-108 in Cell Culture Ap-
proaches to Invertebrate Neuroscience, D. J. Beadle. G. Lees, and S.
B. Kater, eds. Academic Press, London.
Goldberg, J. I., and C. J. Price. 1991. Voltage-gated ionic currents in
neurons isolated from Helisoma trivolvis embryos. In Molluscan
Neurobiology, K. S. Kits. H. H. Boer, and J. Josse. eds. North Holland
Publishing Co.. Amsterdam (in press).
Goldberg, J. I., L. R. Mills, and S. B. Kater. 1991. Novel effects of
serotonin on neurite outgrowth in neurons cultured from embryos
of Helisoma trivolvis. J. Neurobiol. 22: 182-194.
Hadfield, M. G. 1978. Metamorphosis in marine molluscan larvae: an
analysis of stimulus and response. Pp. 165-175 in Settlement and
Metamorphosis oj Marine Invertebrate Lar\ae. F.-S. Chia and M. E.
Rice. eds. Elsevier. New York.
McEuen, F. S. 1985. Reproductive patterns in holothuroids. Ph.D
Thesis. The University of Alberta. Edmonton. Canada.
Morse, D. E. 1990. Recent progress in larval settlement: closing the
gap between molecular biology and ecology. Bull. Mar. Sci. 46(2):
465-483.
Figure 27. Anterolateral ganglion showing sensory neurons (S) with large electron-dense granules (G)
and dendrites (SN) projecting into the epidermis. Bar = 1 /jm.
Figure 28. Higher magnification of an in situ sensory neuron showing characteristic electron-dense
granules (G) interspersed with vesicles (arrows). Bar = 0.5 nm.
Figure 29. Phase-contrast photomicrograph showing an interneuron (N) and a sensory neuron with a
shortened dendritic portion (SD) and a thinner axonal segment (AX ). Note a second unlabelled sensory cell
with a thickened dendritic region. Bar = 5 ^m.
Figure 30. Section through a sensory neuron under in vitro conditions, showing electron-dense granule
(G) and region of vesicles (arrows). Insets: A. Cilium from a dendrite of a sensory neuron. B. Higher mag-
nification of region with vesicles. Bars = 0.5 Mm
76
F.-S. CHIA ET AL.
Morse, D. E., and G. Baxter. 1989. In vitro dissection of chemosen-
sory pathways controlling larval metamorphosis. Am. Zoo/. 29:
30A.
Morse, D. E., H. Duncan, N. Hooker, A. Baloun, and G. Young. 1980.
GABA induces beh;; ••. v.rj] and developmental metamorphosis in
planktonic mollusi i larvae. Fed. Proc. 39: 3237-3241.
Pawlik, J. R. 199(1. Natural and artificial induction of metamorphosis
of Phragmatopoma lapidosa cali/bmica (Polychaeta: Sabellariidae),
with a critical look at the effects of bioactive compounds on marine
invertebrate larvae. Bull. Mar. Sci. 46(2): 512-536.
Richardson, K. C., L. Jarrett, and E. H. Finke. 1960. Embedding in
epoxy resins for ultrathin sectioning in electron microscopy. Slain
Techno/. 35: 313-323.
Schacher, S., and E. Proshansky. 1983. Neurite regeneration by Aplysia
neurons in dissociated cell culture: modulation by Aplysia hemolymph
and the presence of the initial axonal segment. J. Neurosci. 10: 2403-
2413.
Trapido-Rosenthal, H. G., and D. E. Morse. 1986. Availability of
chemosensory receptors is down-regulated by habituation of larvae
to a morphogenetic signal. Proc. Nail. Acad. Sci. USA 83: 7658-
7662.
Wood, R. L., and J. H. Luft. 1965. The influence of buffer systems on
fixation with osmium tetroxide. J. Ultrastruct. Res. 12: 22-45.
Yool, A. J., S. M. Grau, M. G. Hadfield, R. A. Jensen, D. A. Markell, and
D. E. Morse. 1986. Excess potassium induces larval metamorphosis
in four marine invertebrate species. Biol. Bull. 170: 255-266.
Reference: Biol. Bull. 182: 77-96. (February, 1992)
Early Development in the Lancelet (=Amphioxus)
Branchiostoma floridae from Sperm Entry through
Pronuclear Fusion: Presence of Vegetal Pole Plasm
and Lack of Conspicuous Ooplasmic Segregation
LINDA Z. HOLLAND AND NICHOLAS D. HOLLAND
Marine Biology Research Division, Scripps Institution of Oceanography,
University of California San Diego, La Jolla, California 92093-0202
Abstract. Lancelet eggs are described from serial fine
sections before fertilization and at frequent intervals
thereafter until the male and female pronuclei meet at 16
min after insemination. In the unfertilized egg, although
mitochondria, as well as yolk granules, are evenly distrib-
uted (both are absent only from the egg cortex and meiotic
spindle), the mitochondria in the animal third have a more
electron-lucent matrix than those elsewhere. The cortex
of the unfertilized egg is occupied chiefly by cortical gran-
ules, and the subcortical cytoplasm in the vegetal third
includes sheets of dense granules interleaved with cisternae
of endoplasmic reticulum. By 45 s after insemination, ( 1 )
the fertilizing sperm enters (in the animal hemisphere in
three out of three observations), (2) yolk granules become
patchily distributed around the newly entered sperm, (3)
cortical granule exocytosis occurs, and (4) the sheets of
dense granules and associated endoplasmic reticulum ag-
gregate with numerous mitochondria into whorls in a
yolk-free zone near the vegetal pole. These whorls are the
vegetal pole plasm, which is segregated into a single blas-
tomere at each cleavage and might play a role in germ
line determination. By 2 min after insemination, the zone
of cytoplasm near the animal pole with patchily distrib-
uted yolk has enlarged, and the male pronucleus has mi-
grated to the vicinity of the vegetal pole and formed an
aster, at the center of which a few mitochondria are ag-
gregated. In lancelets, unlike ascidians, there is no obvious
widespread ooplasmic segregation or translocation of cy-
toplasm from animal to vegetal pole accompanying the
movement of the sperm. Between 6 and 16 min, (1) the
Received 26 July 1 99 1 ; accepted 31 October 1991.
zone of cytoplasm with patchily distributed yolk enlarges
to occupy about the animal third of the egg, (2) the female
pronucleus forms by fusion of chromosome-containing
vesicles and migrates vegetally, leaving a track of yolk-
poor cytoplasm, and (3) the male pronucleus, surrounded
by increasing numbers of mitochondria, migrates to meet
the female pronucleus just above the equator. In contrast
to current opinion, lancelets differ from ascidians both in
having a vegetal pole plasm and in lacking marked oo-
plasmic segregation.
Introduction
The importance of lancelets in chordate evolution was
first revealed by the embryological studies of Kowalevsky
( 1 865, 1 867). His work stimulated many descriptive stud-
ies on the development of lancelets aimed at clarifying
their phylogenetic relations (e.g., Hatschek, 1882, 1893;
Lwoff, 1892; van Wijhe, 1893; Willey, 1894). In the first
experimental study of this problem, Wilson (1893) in-
vestigated the development of isolated blastomeres and
partial embryos. He concluded that the regulative capacity
of lancelets is intermediate between that of echinoids and
ascidians; that is, blastomere fates become determined in
ascidians, lancelets, and echinoids by the second, third,
and fourth cleavages, respectively.
Until recently, studies of lancelet embryology could deal
only superficially with events before first cleavage because
of difficulty in obtaining the earliest stages: artificial fer-
tilization was never achieved, and, therefore, embryos
were collected after males and females had spawned to-
gether in the field or the laboratory. Thus, descriptions of
early events like the cortical reaction, pronuclear move-
77
78
L. Z. HOLLAND AND N. D. HOLLAND
ments, and maturation divisions (van der Stricht, 1896;
Sobotta, 1895, 1897; Cerfontaine, 1906) were based on
relatively incomplete material.
Conklin ( 1 932, 1 933) reim estigated both the descriptive
and the experimental studies with special attention to
possible similarities between lancelets and ascidians. He
had already established his reputation as an authority on
protochordate development with two papers ( 1 905a. b)
on the embryology of the ascidian Styela partita. His de-
scriptive work ( 1905a) showed that ooplasmic rearrange-
ments between fertilization and pronuclear fusion in -S
partita segregated the following five areas of cytoplasm
destined to be incorporated into specific embryonic tis-
sues: ( 1 ) the yolk-poor, dark yellow myoplasm with abun-
dant mitochondria and pigment granules, destined for the
larval tail muscles, (2) an adjacent light yellow mesen-
chyme material (the myoplasm and mesenchyme together
comprise the mesodermal crescent), (3) the yolk-poor,
clear ectoplasm, the precursor of the ectoderm, (4) the
yolk-rich, dark grey endoplasm, that becomes the endo-
derm, and (5) a light grey cytoplasm destined for the no-
tochord and neural plate. In his experimental work,
Conklin (1905b) reported that individual blastomeres and
groups of blastomeres separated at or beyond the second
cleavage and reared in isolation had the same develop-
mental fate as in the intact embryo. Thus, he concluded
that the ooplasmic segregation created a mosaic of organ-
forming substances in the ascidian embryo and deter-
mined the fate of each region of the uncleaved, fertil-
ized egg.
In 1910, Conklin began to study lancelet development,
but had difficulty obtaining embryos. During the next 22
years, he obtained some additional material, but still
lacked the earliest stages. Therefore, when Conklin finally
published on lancelet embryology, he was forced to rely
on van der Stricht ( 1 896), Sobotta ( 1 897), and Cerfontaine
(1906) for all events before first cleavage. Nevertheless,
Conklin concluded that in regard to pronuclear move-
ments and ooplasmic segregation lancelets were "precisely
like ascidians" (1932) and that "the localizations of ma-
terials in the Amphioxus egg are like those of ascidians,
although not so sharply differentiated" (1933). In other
words, the fate maps of ascidians and lancelets were iden-
tical. Furthermore, in contrast to Wilson ( 1 893), Conklin
(1933) believed that ooplasmic segregation ensured that
"all axes and coles of the future larva are irreversibly de-
termined at or before the first cleavage . . . ," and he
concluded that "development in Amphioxus, as also in
Ascidians, is a mosaic work." This conclusion has been
widely accepted by later biologists (e.g., Brien and Dalcq,
1948; Drach. 1948; Wall, 1990).
It has generally been overlooked that Tung et al. (1958,
1960a, b, 1962a, b) repeated and extended Conklin's ex-
periments on lancelet embryos. They made some changes
in Conklin's fate map, finding in particular that the dis-
tribution of mesodermal material is rather different from
that of ascidians, being more like that of amphibians. In
addition, Tung et al. (1958) supported the view of Wilson
(1893) that the lancelet egg has a considerable regulative
capacity, and they concluded that "the development of
the egg of Amphioxus is, therefore, not a mosaic work as
suggested by Conklin."
In light of the work of Tung et al. (1958), a reinvesti-
gation with transmission electron microscopy (TEM) of
Conklin's descriptive work on early embryology of lan-
celets is especially important. A recent TEM study of lan-
celet development by Hirakow and Kajita (1990) relied
on natural spawnings and thus included only limited ob-
servations on fertilized, uncleaved eggs. This obstacle has
recently been overcome with the development of methods
for spawning and artificially fertilizing lancelet eggs (Hol-
land and Holland, 1989a). In our initial study on the cor-
tical reaction of Branchiostoma floridae, we showed con-
clusively, that unlike ascidian eggs, which lack cortical
granules, lancelet eggs have cortical granules that undergo
exocytosis at fertilization and contribute to the formation
of the fertilization envelope (Holland and Holland,
1989a). In the present work, we extend our fine structural
investigations to cover events between sperm entry and
pronuclear fusion. We have followed the formation of the
pronuclei and pronuclear migrations and have discovered
a conspicuous vegetal pole plasm, but we have found no
evidence for extensive ooplasmic segregation of the as-
cidian type.
Materials and Methods
Specimens of Branchiostoma floridae were collected in
late summer of 1988, 1989, and 1990 in Old Tampa Bay,
Florida. Spawning of females was induced by electrical
shock, and sperm motility was stimulated by 10 mA/
NH4C1 as previously described (Holland and Holland,
1989a). Because only a few of the sperm bound to eggs
undergo the acrosome reaction, it was necessary to use a
concentrated sperm suspension (roughly 1:500 to 1:1000
dilution of dry sperm) to obtain synchronous fertilization.
Development was at 24°C; at that temperature, first
cleavage occurs about 30 min after insemination, and
gastrulation begins at about 5 h.
For TEM. eggs were fixed in 1% K2Cr2O7, 3% glutar-
aldehyde, 0.7 M NaCl pH 7.4, and postfixed in the same
buffer" plus 1% OsO4 and 0.7 Al NaCl (Holland, 1988).
For low-power TEM, some unfertilized eggs and some at
45 s after insemination were fixed as above with the NaCl
lowered to 0.45 A/ to prevent the shrinkage that occurred
before completion of the cortical reaction in eggs fixed in
higher tonicity. Eggs were dehydrated in an ethanol series
and embedded in Spurr's resin. This method, chosen be-
EARLY DEVELOPMENT IN A LANCELET
79
cause of good preservation of organelles, does not preserve
some constituents of the chromosomes, which thus have
a low electron density. We fixed unfertilized eggs as well
as fertilized eggs at 15-s intervals up to 1.5 min after in-
semination, at 30-s intervals up to 2 min after insemi-
nation, at 1-min intervals up to 10 min after insemination,
and at 2-min intervals to 32 min after insemination, the
time of first cleavage. To determine the timing of pro-
nuclear movements, serial 1-2 j/m sections were stained
with 1% toluidine blue in sodium borate and examined
by light microscopy (LM). In the light of those results,
one to five eggs were serially fine-sectioned at each of the
following intervals after insemination: 0 s, 45 s, 2 min, 6
min, 10 min, and 16 min.
Results
Unfertilized egg (Figs. 1. 2)
The spawned, unfertilized egg of Branchiostoma flori-
dae is about 140 /urn in diameter and is arrested in meta-
phase of the second meiotic division. The animal pole is
marked externally by the first polar body and internally
by the second meiotic spindle (Figs. 1 ; 2A, B). Surrounding
the egg and overlying the first polar body is a vitelline
layer (Figs. 1; 2A, C; 3D). The egg cytoplasm contains a
peripheral layer of cortical granules, which are closely ap-
posed to one another except where the meiotic spindle
intervenes (Figs. 1 ; 2B, C). During the first minute after
insemination, the cortical granules undergo exocytosis,
initiating elevation of a fertilization envelope as previously
described (Holland and Holland, 1989a). The first polar
body, of both unfertilized and fertilized eggs (Fig. 3D),
typically includes some unreacted cortical granules and
a cluster of chromosomes; a nucleus is lacking, although
there are frequently a few profiles of nuclear envelope
(Fig. 4B).
Within the egg cytoplasm, yolk granules, 2-5 ^m in
diameter, have a relatively even distribution, being ex-
cluded only from the cortical cytoplasm and the meiotic
spindle (Fig. 2A). In about a 30° arc near the vegetal pole,
just interior to the cortical granules, are several sheets of
dense granular material stacked 2-4 deep, parallel to the
egg plasma membrane (Figs. 1; 2C, D). These sheets are
usually, but not always, interleaved with sheets of smooth
endoplasmic reticulum (SER) (Fig. 2D), which are rare
elsewhere in the cytoplasm. Although the granules com-
prising the sheets are similar in size to ribosomes, the
arrangement of the granules and ER differs from that of
rough ER: between two cisternae of SER there is only
one sheet of granules, and it is separated from the SER
on each side by a space 50 to 75 nm wide. Some mito-
chondria are usually situated near the sheets, but not in
conspicuously greater abundance than elsewhere in the
cytoplasm where they are fairly uniformly distributed
(approximately 30/100 Mm3)-
Although all the mitochondria in the unfertilized egg
are about the same size (0.5 X up to 3 nm), those in a
zone about 35 jum deep around the meiotic spindle have
an electron-lucent matrix (Fig. 2E), while those elsewhere
have a much denser matrix (Fig. 2D). This difference does
not appear to be a fixation artifact since there is a narrow
transition zone where both types of mitochondria co-occur
(Fig. 5C). The distribution of the two types was similar
in all eggs examined and did not change after fertilization,
at least up until formation of the zygote nucleus.
Sperm entry: 30-45 s after insemination (Figs. 1, 3, 4)
Sperm of Branchiostoma floridae have a compact nu-
cleus about 1.5 ^m in diameter, a midpiece (with two
centrioles and one mitochondrion), and a cup-shaped ac-
rosome that can undergo an acrosome reaction producing
a short acrosomal tubule (Holland and Holland, 1989b;
unpub.). To determine the timing of sperm entry, 1-2
nm sections of eggs fixed at 1 5-s intervals after insemi-
nation were examined by light microscopy (LM). In two
eggs fixed at 30 s after insemination, at the beginning of
the cortical reaction, a sperm was seen by LM attached
to the egg surface in the animal hemisphere via a short
fertilization cone (data not shown). However, since con-
densed sperm nuclei are about the same density and size
as yolk granules, no sperm nucleus could subsequently
be detected in 1 pm sections until about 4 min after in-
semination, when it appeared as a clear sphere about 5
nm in diameter in a small yolk-free zone of cytoplasm
near the vegetal pole. Fluorescent DNA-binding dyes also
failed to reveal the newly entered sperm because its nu-
cleus could not be differentiated from those of non-fer-
tilizing sperm bound to the fertilization envelope. A large
excess of sperm is required for synchronous fertilization,
and many remain associated with the fertilization enve-
lope, both at the animal pole and elsewhere, even after
the cortical reaction (Fig. 3A. B, D). Therefore, to detect
the fertilizing sperm just after entry, serial TEM sections
were made through an egg at 45 s after insemination.
Two serial fine sections from the same egg at 45 s after
insemination are shown in Figure 3A, B. The section in
Figure 3 A approximately bisects the egg, and the first polar
body marks the animal pole at the top of the figure. Figure
3B, which is in the same orientation, but about half-way
between the center and the edge of the egg, includes the
fertilizing sperm (arrow). The sperm (shown at higher
magnification in Fig. 3C) has just entered the egg and is
located in the animal hemisphere underneath the egg
plasma membrane about 30° from the animal pole. This
result, plus the two LM observations of fertilization cones
in the animal hemisphere, shows that sperm can fertilize
80
L. Z. HOLLAND AND N. D. HOLLAND
mm
45 sec
2 min
10 min
mm
Figure 1. Diagrams of the unfertilized egg and fertilized eggs through 16 min after insemination. The
distribution of yolk and the sperm aster are not shown. The unfertilized egg has the first polar body and is
in second meiotic metaphase. The egg cortex contains numerous cortical granules, and in the vegetal third
of the egg there are sheets of dense granules interleaved with endoplasmic reticulum just beneath the layer
of cortical granules. At 45 s. most of the cortical granules have undergone exocytosis, the sperm has entered,
the axoneme has largely dispersed, and the sheets of dense granules and endoplasmic reticulum have formed
into whorls to constitute the vegetal pole plasm. By 2 min. the egg is in second meiotic anaphase. the cortical
reaction is complete, and the sperm nucleus has migrated to the vegetal cytoplasm and formed a small aster,
with which a small cluster of mitochondria is associated. By 6 min, the egg is in telophase of the second
meiotic division, the sperm nucleus has swollen, and the peripheral chromatin has condensed more than
the central chromatin. A cloud of mitochondria surrounds the sperm nucleus. By 10 min, the second polar
body has formed. The nuclear envelopes have formed around individual or groups of maternal chromosomes.
These chromosome-containing vesicles are fusing to form the maternal pronucleus. The enlarged male
pronucleus is surrounded by a larger cloud of mitochondria and has migrated partway towards the female
pronucleus. By 16 min. the second polar body has separated from the egg. and the pronuclei have met and
are associated with an asymmetric cloud of mitochondria brought by the male pronucleus.
eggs ofBranchiostomafloridaem the animal hemisphere,
although the sample size is far too small to rule out the
possibility that sperm can also enter in the vegetal hemi-
sphere.
By 45 s after insemination, the cortical reaction is nearly
complete, and only a few unreacted cortical granules re-
main. The yolk granules are still evenly distributed except
in a broad area around the newly entered sperm where
they are somewhat sparser (Fig. 3B). At higher magnifi-
cation, the sperm mitochondrion, one of the two centrioles
(the other is out of the plane of section), and microtubules
of the axoneme are visible in the egg cytoplasm (Fig. 3C).
The nuclear envelope has already disappeared, and the
chromatin has decondensed at the nuclear periphery and
in patches deeper in the nucleus.
The first polar body adheres to the fertilization envelope
as it rises from the egg surface (Fig. 3D). Within the egg,
the meiotic spindle, with chromosomes still aligned on
the metaphase plate, remains associated with relatively
lucent mitochondria and is closely surrounded by yolk
EARLY DEVELOPMENT IN A LANCELET
81
granules (Figs. 3D, 4A). Deeper in the cytoplasm, es-
pecially in the animal hemisphere, the mitochondria are
frequently aggregated into clusters (Fig. 4D). This ar-
rangement of mitochondria persists at least until first
cleavage; there is no apparent movement of mitochondria
from the cortical cytoplasm to the vegetal hemisphere to
surround the sperm nucleus as occurs in ascidians during
ooplasmic segregation. These aggregates of mitochondria
and others described below associated with the pole plasm
and male pronucleus are very small compared to those
in the ascidian myoplasm and are not large enough to be
detected in living eggs with fluorescent mitochondrial
dyes. Thus both DioCifS) and rhodamine 123 seemed to
show a uniform distribution of mitochondria for at least
20 min after insemination. Although the subsequent de-
velopment of eggs in DioC,(3) was not tested, eggs reared
in the dark in rhodamine 123 developed into normal 3-
day larvae.
The subcortical sheets of dense granules and associated
ER in the vegetal third of the unfertilized egg have come
together in a yolk-poor zone of cytoplasm to one side of
the vegetal pole; 6 to 10 layers, each composed of a sheet
of granules and a cisterna of SER, are roughly spiraled
together so that in cross section the pattern resembles that
of a fingerprint (Fig. 4C; for higher magnification see ap-
pearance at 6 min Fig. 7D). At the periphery of these
whorls are numerous mitochondria (Fig. 4C). This reor-
ganization does not appear to be associated with a massive
inflow of materials from other regions of the egg. Because
of its location, we will call this specialized region of cy-
toplasm the vegetal pole plasm. From this point in de-
velopment, the appearance of these whorls remains rel-
atively constant, at least through formation of the zygote
nucleus. The pole plasm is visible in toluidine blue-
stained, 2 nm sections as reddish-purple strands in a yolk-
free zone near the vegetal pole. At each cleavage, at least
through the early blastula, it is segregated into a single
blastomere (data not shown).
Sperm pronucleus near vegetal pole: 2-6 min after
insemination (Figs. 1, 5)
By 2 min after insemination, the male pronucleus, re-
gardless of the point of sperm entry, is located in the egg
cortex near the vegetal pole. In TEM sections of an egg
at 2 and one at 6 min after insemination (Figs. 5A, 6A),
and in LM sections through ten eggs at 3 to 6 min after
insemination, the male pronucleus was always near the
vegetal pole. These results are consistent with previous
LM studies demonstrating that the swollen male pronu-
cleus first becomes visible near the vegetal pole of the egg
of Branchiostoma lanceo/atiim (van der Stricht, 1985;
Sobotta, 1897; Cerfontaine, 1906). Presumably, as in
ascidian eggs (Speksnijder et al.. 1989), sperm entering
the animal hemisphere of the ascidian egg are rapidly
translocated close to the vegetal pole. At 2 min after
insemination, the male pronucleus is about 2.5 ^m in
diameter and is less dense than the cytoplasm (Figs. 5A,
B). There is no trace of a nuclear envelope surrounding
the male chromatin (Fig. 5B, insert). A sperm aster is
present near the male pronucleus (Fig. 5D) (none of our
sections went through both the male pronucleus and the
aster). We did not see the male centrioles at this time;
presumably the sections containing them were lost. The
cytoplasm around the male pronucleus is not enriched in
mitochondria, as it is in ascidians, although a few mito-
chondria are aggregated at the convergence of the astral
rays (Fig. 5D).
In the cytoplasm, yolk granules are somewhat less nu-
merous in the animal hemisphere than in the vegetal
hemisphere (Fig. 5 A). This is the opposite of the situation
in ascidians where the yolk is concentrated in the animal
hemisphere during the first phase of ooplasmic segrega-
tion. The meiotic spindle is still present at the animal
pole, but the female chromosomes, which are inconspic-
uous, have begun to migrate to opposite poles of the spin-
dle (data not shown). The pole plasm with its sheets of
dense granular material is unchanged from 45 s after in-
semination (see appearance at 6 min. Figs. 7A, B).
Beginning of pronuclear migration: 6-10 min after
insemination (Figs. 1, 6, 7)
By 6 min after insemination, the male pronucleus,
which has enlarged to 5 /urn in diameter and developed
an irregularly lobed outline, has migrated from close to
the egg cortex about 30 ^m towards the center of the egg
(Figs. 1, 6A). The chromatin of the male pronucleus re-
mains decondensed at the center, but has become more
condensed in the peripheral lobes and is partly bounded
by a nuclear envelope (Fig. 7F. arrow). Numerous mi-
tochondria, apparently recruited from the cytoplasm in
the vegetal hemisphere, closely surround the periphery of
the male pronucleus (Fig. 7B). Although the sperm aster
was not seen at this stage, as in the preceding stage (see
Fig. 5D) it is likely that the aggregate of mitochondria
also converges upon the sperm aster.
The maternal chromosomes have moved apart on the
meiotic spindle. Those destined to form the female pro-
nucleus are rounded, about 6 /im in diameter, and lie in
a yolk-poor zone slightly away from, and to one side of,
the animal pole (Figs. 6A, C; 7A). A few fragments of
nuclear envelope have formed at the periphery of the
chromosomes, which are still associated with spindle fibers
(Fig. 7A). At the animal pole, there is a bulge in the egg
surface, evidently the beginning of the second polar body
(Fig. 6D). From previous observations, this polar body
forms about 8 min after insemination (Holland and Hoi-
82
L. Z. HOLLAND AND N. D. HOLLAND
.
. ••^2,,::V' '..
*'-'1i}-yf' • • ....•"•"••*/.-' '
r-:
^.-,., ' :-. •• "
EARLY DEVELOPMENT IN A LANCELET
83
land, 1989a). As in the preceding stage, yolk granules are
somewhat scarcer in the animal hemisphere (Fig. 6A),
and the pole plasm is prominent near the vegetal pole
(Figs. 6B; 7C. D).
Formation of the Jem ale pronucleus and migration of
pronuclei: 10-16 min after insemination (Figs. 1, 8, 9)
By 10 min after fertilization, the male pronucleus has
continued its migration towards the female pronucleus
and enlarged to about 6 ^m in diameter. (Fig. 8B. C). The
chromatin is of uniform density, similar to that of the
peripheral lobes at 6 min after insemination, and the nu-
cleus is bounded by a nuclear envelope (Fig. 8C). The
area of yolk-free cytoplasm around the male pronucleus
has enlarged to about 1 5 /^m in diameter and has become
rich in mitochondria; there are about four mitochondria
per /iirr — approximately ten times the concentration
elsewhere in the cytoplasm.
The second polar body has formed, but is still attached
to the egg (Figs. 8A, 9D). At the point of attachment,
there is a prominent density (Zwischenkorper) through
which the spindle microtubules pass (Fig. 9D, insert). In
the polar body, there are several chromosome-containing
vesicles, each bounded by its own nuclear envelope
(Fig. 9D).
Within the egg, individual female chromosomes or
groups of chromosomes have become surrounded by nu-
clear envelopes. Depending on the egg, these chromo-
some-containing vesicles are either in the process of fusing
and are still located close to the animal pole (Fig. 9A-C)
or they have completed fusing into a single female pro-
nucleus 5.5 X 8 nm in diameter, which has migrated
somewhat off-center to just above the equator (Figs. 1,
8B, D). The female pronucleus does not migrate further.
The nuclear matrix is of low electron density but contains
small scattered areas of higher density (Figs. 8D; 9B. C).
The female pronucleus is in a large, irregular area of yolk-
poor cytoplasm, which extends to the yolk-poor cytoplasm
near the animal pole (Fig. 8B). Although mitochondria
are not uncommon near the female pronucleus, they are
four to five times less numerous than those around the
male pronucleus (compare Fig. 8D and C).
The pole plasm (Figs. 1; 8A; 9E, F) is little changed
from earlier times. However, in some places, the strands
of dense material are no longer closely associated with
ER and have lost their parallel relation to one another
(Fig. 9E, F).
Pronuclear fusion: 16 min after insemination
(Figs. 1, 10)
By 16 min after insemination, the male pronucleus has
migrated to, and fused with, the female pronucleus (Figs.
1 : 10A, E, F). The resulting zygote nucleus, which is about
8X12 /urn in diameter, lies in a zone relatively free of
yolk about 17 //m in diameter in the animal hemisphere
just above the equator, about half-way between the edge
and the center of the egg. At one side of the nucleus,
presumably that deriving from the male pronucleus, is a
large aggregate of mitochondria (Fig. 10E). Thus a mi-
tochondria-rich zone of cytoplasm surrounds the newly
formed zygote nucleus in both lancelets and ascidians.
However, this zone is vastly larger in ascidians. and comes
not from the vegetal cytoplasm as the male pronucleus
migrates through it, but from the cortical cytoplasm,
which collects around the male pronucleus during oo-
plasmic segregation (Zalokar and Sardet, 1984).
The zygote nucleus contains a few small, dense inclu-
sions like those previously described in the female pro-
nucleus (Fig. 1 OF) and is bounded everywhere by a nuclear
envelope (Fig. 10E, F). Although some microtubules, ev-
idently part of the astral rays, were seen near the nucleus,
no centrioles were encountered in our sections.
By 1 6 min, the second polar body has separated from
the egg, but their plasma membranes remain closely ap-
posed (Fig. 10B-D). The second polar body tends to re-
main at the animal pole during the cleavage stages, al-
though Hirakow and Kajita (1991) sometimes observed
it in other locations. Within the polar body, the chro-
mosome-containing vesicles have fused into a single nu-
cleus.
Discussion
Previous work on the early embryology of lancelets has
been largely on the European species, Branchiostoma lan-
ceolatum (Wilson, 1893; van der Stricht, 1896; Sobotta,
1897;Cerfontaine, 1906; Conklin, 1932. 1933), and to a
Figure 2. TEMs of unfertilized eggs with animal pole uppermost. A. Central section through the first
polar body at top adjacent a small yolk-free area including the second meiotic spindle. At this magnification,
the spindle fibers and chromosomes cannot be resolved. Scale bar: 20 jim. B. Higher magnification of the
second meiotic spindle The first polar body is not in the plane of section. Mitochondria near the meiotic
spindle generally have an electron lucent matrix (arrow). Scale bar: 2 ^m. C. The vegetal pole. Parallel sheets
of dense granules interleaved with endoplasmic reticulum (arrows) lie just beneath the layer of cortical
granules. Scale bar: 2 ^m. D. Higher magnification of the sheets of dense granules and associated endoplasmic
reticulum at the vegetal pole. Mitochondria (M) have a relatively electron-dense matrix. Scale bar: 0.5 ^m
E. Higher magnification of the electron lucent mitochondria (M) near the animal pole. Scale bar: 0.5 /jm.
Chromosomes (CH), cortical granule (CG). vitelline layer (VL). yolk granule (YG).
•« . • • .- • ••
-
. ' ' :: V, •';" . ..
:
.
HL
*$!&:•."'•. ,'• .'^"":
Figure 3. TEMs of eggs at 45 s after insemination. A. Section through the first polar body (PB) at top.
The cortical reaction is in progress. Many supernumerary sperm (arrows) are associated with the rising
fertilization envelope. The plane of section does not pass through the meiotic spindle. Scale bar: 20 ^m. B.
Section through the same egg as in (A) in the same orientation about halfway between the center of the egg
and the periphery. The fertilizing sperm (arrow) has entered into the animal hemisphere, and the yolk is
patchy in its neighborhood. Scale bar: 20 /jm. C. Higher magnification of the fertilizing sperm in (B). The
nucleus is associated with the sperm mitochondrion (M) and centriole (C). The nuclear envelope has dis-
appeared and the chromatin has begun to decondense. The axoneme (A) has largely dispersed. Scale bar:
0.5 ^m. D. The first polar body, which contains several unreacted cortical granules (CG), and whorls of
nuclear envelope (arrow) is sandwiched between the vitelline layer (VL) and material derived from the
cortical granules, the hyaline layer (HL). Yolk granules closely surround the meiotic spindle (MS). Scale
bar: 2.0 ^m.
84
EARLY DEVELOPMENT IN A LANCELET
85
1 * *i»ji
Figure 4. TEMs of eggs at 45 s after insemination. A. Higher magnification of chromosomes (CH) on
the meiotic spindle. Scale bar: 0.5 ^m. B. Higher magnification of the polar body in 3D showing the whorl
of nuclear envelope (arrow). Scale bar: 0.5 ^m. C. The sheets of dense granules, endoplasmic reticulum,
and mitochondria (M) that constitute the vegetal pole plasm. Scale bar: 2 pm. D. Patchy yolk distribution
and aggregated mitochondria (arrows) in the animal hemisphere. Scale bar: 1 ^m.
86
L. Z. HOLLAND AND N. D. HOLLAND
Figure 5. TEMs of eggs at 2 min after insemination. A. Cross section through meiotic spindle (top) and
the nucleus of the fertilizing sperm (arrow). The cortical reaction is complete. The indentation in the egg at
the animal pole is probably an artifact due to the hypertonicity of the fixative. The polar body and fertilization
envelope are not in the figure. Scale bar: 20 ^m. B. Higher magnification of the fertilizing sperm nucleus in
(A). The aster is out of the plane of section. Scale bar: 1 fim. (B. inset) The edge of the male nucleus at
higher magnification. There is no nuclear envelope. Scale bar: 0.5 ^m. C. Co-occurrence of mitochondria
with dense matrix (DM) and lucent matrix (LM) near the animal pole. Scale bar: 0.5 ^m. D. The aster
associated with the male nucleus in (D). There is a small cluster of mitochondria where the astral microtubules
(MT) converge (asterisk). The male nucleus and centrioles are out of the plane of section. Scale bar: 0.5 nm.
EARLY DEVELOPMENT IN A LANCELET
87
Figure 6. TEMs of eggs fixed at 6 min after insemination. A. Section through the male pronucleus
(arrow). Animal hemisphere is uppermost. The section does not pass through the meiotic spindle and polar
body. The female chromosomes are located about 1 1 o'clock near the animal pole (asterisks), but are not
visible at this magnification due to their low contrast. Scale bar: 20 ^m. B. Section through the same egg as
in (A) that passes through the vegetal pole plasm (PP). The indentation at the vegetal pole is probably an
artifact due to the high tonicity of the fixative, but marks the site of the future cleavage furrow. Scale bar:
20 Mm. C. Three female chromosomes from the same egg as in A and B. Scale bar: 4 Mm. D. Bulge at the
animal pole at the site of formation of the second polar body. The polar body chromosomes are not in the
plane of section. Scale bar: 2 ^m.
lesser extent, on the Asian species, B. belcheri (Tung et
ai. 1958, 1960a, b. 1962a, b; Hirakow and Kajita, 1990,
1991). There are no marked differences between these
species. Thus, although aside from a few micrographs of
Hirakow and Kajita (1990), our work on B floridae is
the only TEM study on the earliest embryonic stages, and
it is likely that our results also apply to other species of
Branchiostoma; the largest ovarian oocytes have virtually
the same fine structure in B. floridae. B. lanceolatum, and
B. belcheri (reviewed in Holland and Holland. 1991), as
do the blastomeres ofB.JIoridae and B. belcheri (Hirakow
and Kajita 1990, 1991; Holland and Holland, unpub.).
Position of sperm entry; formation and migration
of the pronuclei
Sobotta ( 1 897) depicted a sperm entering a lancelet egg
with its tail extending into the perivitelline space and its
nucleus with the same size and staining properties as a
yolk granule. He maintained that, although sperm can
enter the egg anywhere on the surface, they usually do so
near the vegetal pole. He, along with van der Stricht (1896)
and Cerfontaine ( 1 906), thought that the fertilizing sperm
first developed into a dark-staining irregular mass near
the vegetal pole before swelling into a clear, spherical pro-
nucleus. This observation led van der Stricht ( 1 896), Cer-
M
Figure 7. TEMs of eggs at 6 min after insemination. A. Higher magnification of the female chromosome
at the far right in Figure 6C. There is no nuclear envelope. The chromosome is still associated with microtubules
of the meiotic spindle (MT). Scale bar: 1 /jm. B. Higher magnification of the male pronucleus in Figure 6A.
It is closely surrounded by mitochondria, and a partial nuclear envelope has formed (arrows). The aster is
not in the plane of section. Scale bar: 0.5 nm. C. Higher magnification of the vegetal pole plasm in Figure
6B. Numerous mitochondria (M) are associated with the sheets of dense granules. Scale bar: 2 tim. D. High
magnification of the vegetal pole plasm in C. Mitochondria (Ml are closely associated with sheets of endo-
plasmic reticulum (ER) that lie in between the sheets of dense granules. Scale bar: 0.2 urn.
88
EARLY DEVELOPMENT IN A LANCELET
89
Figure 8. TEMs of eggs at 10 min after insemination. A. Section through the second polar body (PB),
male pronucleus (arrow), and vegetal pole plasm (PP). Black dots in the yolk-free zone around the male
pronucleus are mitochondria. The yolk is patchily distributed in the animal hemisphere. Scale bar: 20 yum.
B. Section through the same egg as in A, about 10 Mm deeper, which passes through the female pronucleus
(double arrow). The yolk-free patch of cytoplasm underlying the male pronucleus is at lower right (single
arrow). Scale bar: 20 pm. C. Higher magnification of the male pronucleus and its cloud of mitochondria in
A. The nuclear envelope is complete except in a few spots (arrows). Scale bar: 1 pm. D. Higher magnification
of the female pronucleus in B. The nuclear envelope is complete except in a few areas (arrows). The nucleus
contains a few dense patches. Scale bar: 1 urn.
fontaine (1906), and Conklin (1932) to conclude that the
sperm always enters near the vegetal pole. In contrast,
our results show that the fertilizing sperm can enter the
egg near the animal pole (three out of three observations).
although more extensive study would be required to show
whether they always do so. The axoneme enters with the
nucleus, mitochondrion, and centrioles, but rapidly dis-
appears. Then the sperm nucleus undergoes two phases
90
L. Z. HOLLAND AND N. D. HOLLAND
Figure 9. TEMs of eggs at 10 min after insemination. A. Section near the animal pole with three maternal
chromosome-containing vesicles (CV). Scale bar 4.5 /^m. B. Higher magnification of two of the maternal chro-
mosome-containing vesicles in A. which have begun to fuse at arrows. Scale bar: 0.5 ^m. C. Two maternal
chromosome-containing vesicles that have fused and are connected by a broad bridge. Scale ban 0.5 nm. D.
Higher magnification of the second polar body in Figure 8A. Three chromosome-containing vesicles (CV) are
visible within it. Where it is connected to the egg there is a dense Zwischenkorper (arrow). Scale bar: 1 jjm. (Insert)
The Zwischenkorper at higher magnification showing the microtubules (arrow) remaining from the meiotic spindle.
Scale bar 0.5 ^m. E. Sheets of dense granules and associated mitochondria in the vegetal pole plasm that are no
longer associated with endoplasmic reticulum. Scale bar: 0.5 ^m. F. Sheets of dense granules and mitochondria
in the vegetal pole plasm in relatively close association with smooth endoplasmic reticulum in some places (twin
arrow), but not in others (single arrow). Scale bar: 0.3 /itn.
of migration. First, between 45 s and 2 min after insem-
ination, the male pronucleus evidently migrates rapidly
to the vicinity of the vegetal pole. Second, between 6 min
and 16 min, the male pronucleus migrates slowly back
into the animal hemisphere to meet the female pronu-
cleus.
Soon after entering the egg, the sperm nucleus rapidly
decondenses, staining less intensely with toluidine blue.
Figure 10. LM (A)and TEMs(B-F)ofeggsat 16 min after insemination. A. Section through the animal
pole (top) and zygote nucleus (arrow). The yolk remains patchy near the animal pole. Scale bar: 20 ^m. B.
The proximal pan of the second polar bod\ (PB), which has detached from the egg. Scale bar: 5 fim. C. The
polar body and egg, although no longer in cytoplasmic continuity, remain very tightly apposed (arrows).
Scale bar: 1 nm. D. The distal part of the second polar body. Some of the chromosome-containing vesicles
have fused into a nucleus (N). Scale bar: 1 ^m. E. The portion of the zygote nucleus probably derived from
the male pronucleus has a cloud of mitochondria at one side. Scale bar: 12 ^m. F. A portion of the zygote
nucleus probably derived from the female pronucleus. There are dense patches within the nucleus and the
nuclear envelope is continuous. Scale bar: 0.5 ^m.
41
92
L. Z. HOLLAND AND N. D. HOLLAND
and cannot be seen by LM until 4-5 min after insemi-
nation, when it has swollen considerably. No part of the
fertilizing sperm ever becomes the large, dark-staining ir-
regular structure reported by earlier embryologists — this
structure is clearly not the sperm at all, but the vegetal
pole plasm. The sheets of dense granules belonging to the
vegetal pole plasm are certainly responsible for the mis-
taken view of van der Stricht (1896) and Cerfontaine
(1906) that the sperm tail enters along with the head and
remains behind near the vegetal pole as a sperm remnant
after the male pronucleus swells, develops an aster, and
begins its slow migration. Apparently, both Sobotta ( 1 897)
and Conklin (1932) overlooked the vegetal pole plasm in
eggs with large pronuclei and thus were spared the diffi-
culty of having to explain its presence. The female pro-
nucleus and the swollen male pronucleus are readily vis-
ible by LM, and their migrations were correctly described
by Sobotta (1895,1897), van der Stricht ( 1 896), and Cer-
fontaine (1906).
The second phase of male pronuclear migration begins
just before the second polar body forms. The female chro-
mosomes then move to one side of the animal-vegetal
axis and join to form a female pronucleus, which migrates
to just above the equator to be met by the male pronu-
cleus. Cerfontaine (1906) and Conklin (1932) believed
that, as in ascidians, the site where the pronuclei meet is
the posterior region of the future embryo. However, in
the absence of obvious cytoplasmic markers of either the
posterior or anterior poles, it is puzzling how they could
make the distinction except by analogy with ascidians,
some species of which have yellow pigment granules lo-
calized in the egg before cleavage at the posterior pole of
eggs and embryos (Conklin, 1905a). Conklin (1932)
thought he could distinguish a similar, although less con-
spicuous, marker of the posterior pole in lancelets; how-
ever, as discussed below, the existence of such a marker
is most unlikely.
Pronuclear migrations in lancelets and ascidians are
similar, although some of the details vary. In ascidians
(Conklin, 1905a), as in lancelets, it was formerly believed
that the sperm enters near the vegetal pole. However, it
has since been shown that ascidian sperm can fuse with
all regions of the egg plasma membrane (Ortolani, 1958;
Talevi and Dale, 1986), but preferentially enter the animal
hemisphere (Speksnijder et ai. 1989). In ascidians, as in
lancelets, there are two phases of sperm migration. First,
the sperm is rapidly transported close to the vegetal pole.
Staining with DNA-specinc dyes and an anti-tubulin an-
tibody has shown that during this phase, the ascidian
sperm nucleus remains condensed and is accompanied
by the axoneme (Sawada and Schatten, 1988) — in con-
trast, as we have demonstrated, soon after entering the
egg, the lancelet sperm nucleus decondenses and the ax-
oneme disappears. In both ascidians and lancelets, the
male pronucleus, once in the vegetal cytoplasm, swells,
develops a large aster, and then, in a second slower phase
of migration, moves towards the animal pole. The pro-
nuclei meet just below the equator in ascidians and just
above it in lancelets. An aggregation of mitochondria ac-
companies the male pronucleus in this migration. How-
ever, the mitochondria are far more numerous in ascidians
than in lancelets and are derived, not by gradual recruit-
ment from the surrounding cytoplasm, but from the mi-
tochondria-rich cortical cytoplasm, which flows along with
the male pronucleus from the animal hemisphere to the
vegetal pole and then to the posterior pole to form the
myoplasmic crescent (Sawada and Schatten, 1988; Spek-
snijder et al.. 1989).
The mechanism for migration of the pronuclei in lan-
celets is unclear. In ascidians, the first phase of male pron-
uclear migration occurs concomitantly with a dramatic
shape change and segregation of ooplasm (Jeffery, 1984),
all of which are inhibited by cytochalasin and are thus
probably mediated by the contraction of cortical micro-
filaments (Sawada, 1988; Sardet et al., 1989). We did not
test whether cytochalasin could prevent the first phase of
sperm migration in the lancelet; however, lancelet eggs
undergo neither a shape change (Holland and Holland,
1989a) nor obvious ooplasmic segregation.
In ascidians, the sperm aster is necessary for the second
phase of migration of the male pronucleus and for the
movement of the mitochondria-rich myoplasm from the
vegetal pole towards the posterior pole; both movements
are prevented by agents that disrupt microtubules (Manes
and Barbieri, 1977; Sawada and Schatten, 1988). Whether
microfilaments are also involved is not known. However,
in sea urchins, migration of the male pronucleus, which
also depends on microtubules, is independent of micro-
filaments (Schatten and Schatten, 198 1 ). Thus, in lancelets
as well, although microtubule inhibitors have not been
tested, the sperm aster is probably necessary for migration
of the male pronucleus; the sperm aster is also probably
responsible for the aggregation of mitochondria around
the male pronucleus. Mitochondria do not aggregate
around the female pronucleus, which lacks an aster, or
around the male pronucleus before the aster forms. In
addition, in somatic cells, mitochondria are frequently
seen in close association with microtubules (Heggeness et
al., 1978), which have been shown to function as tracks
for the movement of organelles, particles, and molecules
in somatic cells, eggs, and embryos (Schliwa, 1984; Vale
et al., 1985; Hamaguchi et al., 1986; Kobayakawa, 1988;
Ressom and Dixon, 1988; Yisraeli et al.. 1989).
Vegetal pole plasm
As mentioned above, the vegetal pole plasm was seen
with LM in lancelet eggs but misidentified as the fertilizing
EARLY DEVELOPMENT IN A LANCELET
93
sperm (van der Stricht, 1896; Sobotta, 1897; and Cerfon-
taine, 1906). With TEM, Hirakow and Kajita (1990) il-
lustrated the pole plasm in fertilized, uncleaved eggs in
their figure 12, but interpreted it as an "unusual stack of
rough endoplasmic reticulum rarely encountered."
The vegetal pole plasm of lancelets has the components
typical of pole plasms in other organisms, i.e., numerous
mitochondria, conspicuous aggregates of dense fibro-
granular material, and profiles of endoplasmic reticulum.
The precise configuration of the pole plasm in lancelets,
however, has not been described in any other organism.
Among the deuterostomes, vegetal pole plasm has been
seen only in chaetognaths and anuran amphibians. It does
not occur in appendicularian (Holland et al. 1988) or
ascidian tunicates, or in echinoderms; its possible presence
in hemichordates has not been investigated (Eddy, 1975).
In many organisms, the vegetal pole plasm is destined to
be included in the primordial germ cells and, thus, is
termed "germ plasm." The germ plasm is enriched in
RNA, and some mRNAs and proteins specific to it have
been identified (Phillips, 1982, 1985; Yamaguchi et al,
1983; Hay et al, 1988; Nakazato and Ikenishi, 1989).
Nevertheless, it is not known how the germ plasm acts in
germ cell determination for any animal (Davidson, 1986).
The germ cells are typically endodermal derivatives in
animals with germ plasm, e.g., chaetognaths and anurans,
but are usually mesodermal derivatives in those lacking
germ plasm, e.g.. urodele amphibians and probably as-
cidians (Nieuwkoop and Sutasurya, 1976, 1979; Nieuw-
koop, 1991). For the lancelet Branchiostoma belcheri at
the 32 cell-stage, the most vegetal tier of blastomeres, one
of which presumably contains the vegetal pole plasm, is
destined to form endodermal structures such as the gut;
embryos lacking these blastomeres rarely form endoder-
mal structures (Tung et al., 1960a). Thus, the vegetal pole
plasm of lancelets may be included in endodermal cells,
and the germ cells would thus be expected to be endo-
dermal in origin. In B lanceolatum, Boveri (1892) found
primordial germ cells in segmentally arranged outpock-
etings of the myocoel in relatively late larvae, which sug-
gested to Nieuwkoop and Sutasurya (1979) that the germ
cells would be mesodermal, not endodermal, derivatives.
However, Boveri lacked earlier larvae and thus could not
have determined if the germ cells had arisen in the my-
ocoel or migrated from elsewhere. The possibility that the
vegetal pole plasm in lancelets is incorporated into the
germ cells merits investigation. The first two blastomeres,
when separated, can each give rise to a normal larva (Wil-
son, 1893; Conklin, 1933; Tung et al.. 1958); however,
no such embryo has been raised long enough to determine
if gonads formed.
The dense granular material in vegetal pole plasm or
germ plasm is thought to be related to, and possibly de-
rived from, nuage — dense fibrogranular aggregates con-
taining protein or RNA that frequently occur in associ-
ation with mitochondria near the nucleus of growing oo-
cytes. Nuage occurs in lancelets (Guraya, 1983; Aizenstadt
and Gabaeva, 1987; Holland and Holland, 1991) and in
most other organisms that have germ plasm. Nuage is
also present in many animals lacking germ plasm, in-
cluding echinoderms, and among the chordates, ascidians,
reptiles, and birds (Eddy, 1975). In the European lancelet
Branchiostoma lanceolatum, Guraya (1968, 1979) found
that nuage contained protein, lipoprotein, and RNA. In
mid-oogenesis, the aggregates of nuage break up and are
distributed throughout the cytoplasm, becoming localized
in the cytoplasm at the vegetal pole in the largest oocytes
(Guraya, 1983). At least part of the nuage may be the
source of the sheets of dense aggregates present just interior
to the cortical granules in B.floridae, which coalesce after
insemination to form the vegetal pole plasm.
Cytoplasmic specializations at the animal pole
Eggs of Branchiostoma floridae have two specializations
near the animal pole: first, in both unfertilized and fer-
tilized eggs, animal pole mitochondria are relatively elec-
tron-lucent, and second, in fertilized eggs, the yolk be-
comes patchy in the animal hemisphere. In blastulae of
axolotls, there is a similar animal-vegetal difference in
mitochondria; those in animal cells are larger and have a
much less dense matrix than those of vegetal pole cells
(Nelson et al., 1982).
Relatively yolk-poor areas at the animal pole (namely,
animal pole plasms) have been described in fertilized eggs
of both invertebrates and vertebrates, for example, oli-
gochaetes (Shimizu, 1989), lampreys (Nicander et al.,
1968), and amphibians (Wakahara, 1989). Typically, these
areas appear either during the meiotic divisions or soon
after fertilization. In ascidians, after germinal vesicle
breakdown, the material from the germinal vesicle be-
comes localized at the animal pole. Following fertilization,
this yolk-poor cytoplasm follows the myoplasm to the
vegetal hemisphere and then migrates with the male pro-
nucleus back towards the animal hemisphere, finally
coming to surround the zygote nucleus (Conklin, 1905a,
b; Jeffery, 1984). In contrast, in lancelets, the yolk-free
patches at the animal hemisphere do not appear to derive
from the germinal vesicle. Cerfontaine (1906) mistakenly
depicted the remnant of the germinal vesicle persisting to
one side of the meiotic spindle; when the germinal vesicle
breaks down, however, its substance rapidly blends with
the cytoplasm, and the yolk becomes uniformly distrib-
uted except immediately around the second meiotic spin-
dle (Conklin, 1933; the present work).
The yolk-free patches that develop after fertilization in
the animal hemisphere near the newly entered sperm do
not follow it to the vegetal pole. Instead, a yolk-free area
94
L. Z. HOLLAND AND N. D. HOLLAND
forms de novo around the male pronucleus as the second
phase of migration begins. The yolk-free patches near the
animal pole may form either by one or more of the fol-
lowing: ( 1) an expansion of the egg volume at the animal
pole, or (2) an aggregation and movement of yolk towards
the vegetal pole, or (3) a movement of yolk-free cytoplasm
to the animal pole. The last explanation is perhaps the
most likely because in many eggs (e.g.. barnacles, oligo-
chaetes, lampreys, and teleost fish) there is such a flow of
cytoplasm to the animal pole from the interior or from
the peripheral layers of the egg (reviewed by Wall, 1990).
The animal cytoplasm has been studied much less than
the vegetal cytoplasm in regard to its role in embryogen-
esis. In general, the animal cytoplasm is destined to form
ectoderm. In lancelets, the most animal of the four tiers
of blastomeres at the 32-cell stage, if isolated, forms only
epidermal structures; however, removal of this tier does
not affect the normal development of the larva (Tung el
al., 1960a). Thus, while destined to form ectoderm, there
are no substances unique to this layer that cannot be du-
plicated by other blastomeres.
Ooplasmic segregation
Since the work of Conklin (1932, 1933), it has been
generally believed that ooplasmic segregation occurs in
lancelets exactly as in ascidians (e.g.. Brien and Dalcq,
1948; Drach, 1948). Conklin maintained that "the local-
izations of materials in the Amphioxus egg are like those
of ascidians, although not so sharply differentiated."
Lacking the stages before the pronuclei meet, he found
evidence for ooplasmic segregation in lancelets in the fig-
ures of Sobotta ( 1 896) and van der Stricht ( 1 897), although
they made no such claims. In addition, Conklin (1933)
was convinced from his own sections of eggs just before
first cleavage that the mesodermal and chorda-neural
crescents were distinguishable from endodermal and ec-
todermal areas; the mesodermal crescent was particularly
visible because it consisted of more deeply staining cy-
toplasm. Curiously, Conklin did not mention the patch-
iness of yolk at the animal pole, although this was shown
by Cerfontaine(1906).
Neither our results, nor those of Hirakow and Kajita
( 1 990), have revealed any evidence for ooplasmic segre-
gation in Branchiostoma such as occurs in ascidians. In
lancelets, there are some small, localized aggregations of
mitochondria in the animal cytoplasm, but no apparent
segregation of mitochondria to the vegetal cytoplasm. The
mitochondria associated with the zygote nucleus appear
to be recruited by the migrating sperm nucleus and do
not derive from the peripheral cytoplasm. Nowhere else
in the fertilized lancelet egg is there a large concentration
of mitochondria, comparable to that in the myoplasm of
ascidian eggs. In addition, we found no differences in
staining between regions destined to become the meso-
derm, notochord, neural plate, or endoderm. As our Fig-
ures 8 A, B, and 10A show, there is no crescent of more
deeply staining cytoplasm anywhere in the egg. In the
lancelet egg, the only type of dense granule is the yolk
granule. The only regional difference in yolk is that it is
scarcer in the animal hemisphere, destined for ectoderm
and neural plate, than in the vegetal hemisphere, destined
for mesoderm, endoderm, and notochord. It is very un-
likely that the discrepancies between Conklin's conclu-
sions and our finding are due to species differences. Al-
though no TEM of fertilized eggs of B. lanceolatum has
been done, yolk granules in the unfertilized egg are iden-
tical to those from B floridae (Holland and Holland,
1991).
Conklin's interpretations of his own sections and the
figures of others appear to have been chiefly based on his
own preconceptions. In his ascidian paper ( 1 905a), before
obtaining any lancelet embryos, Conklin had concluded
from van der Stricht (1896) that lancelet eggs must un-
dergo ooplasmic segregation as in ascidians. Subsequently,
Conklin interpreted all the evidence to support his ideas.
First, he erroneously inferred from Sobotta (1897), who
illustrated vesicles in the egg cortex, that the cortex, like
that of ascidian eggs, contained not cortical granules, but
numerous mitochondria. On the contrary, TEM has
demonstrated that the cortex of lancelet eggs is packed
with cortical granules and is nearly free of mitochondria
(Holland and Holland, 1989a).
Next, Conklin was led further astray by errors of van
der Stricht (1896), who thought that cortical granule dis-
appearance resulted from their migration into the interior
of the egg where they became the yolk-free patches in the
animal cytoplasm. Conklin decided this was wrong in part;
on disappearing, the peripheral cytoplasm migrated not
into the interior of the egg, but to the vegetal pole and
thus was the equivalent of the ascidian myoplasm. In
truth, at fertilization, the cortical granules disappear be-
cause they undergo exocytosis and contribute to the for-
mation of the fertilization envelope (Sobotta, 1897; Hol-
land and Holland, 1989a). As further evidence of my-
oplasm, Conklin cited van der Stricht's figure 12, which,
in fact, shows the yolk-free cytoplasm surrounding the
vegetal pole plasm, which van der Stricht thought was the
sperm remnant.
Finally, when interpreting his own sections, Conklin
apparently saw things that simply weren't there. Publish-
ing before it was common practice to photograph LM
sections, Conklin ( 1933) drew "actual sections" (his text
figure A) of fertilized Branchiostoma eggs showing the
ectoplasm, the chorda-neural crescent, the endoderm, and
the mesodermal crescent with its distinctive granules. This
erroneous drawing has been reproduced in modern texts
EARLY DEVELOPMENT IN A LANCELET
95
(e.g., Wickstead, 1975) and is the sole evidence for an
ascidian-like ooplasmic segregation in lancelets.
Ooplasmic segregation and chordate phytogeny
Discussions about the phylogenetic origin of the chor-
dates and the arrangements of the chordate subphyla are
still highly contentious (<.;/.' Ghiselin el a!., 1986; Jefferies,
1986; Erwin, 1991). The present study has demonstrated
that a conspicuous, ascidian-style ooplasmic segregation
does not occur in acraniates. Importantly, we have shown
that such segregation is not a synapomorphy of ascidians
and acraniates; instead it may be no more than an auta-
pomorphy of ascidian tunicates, because conspicuous cy-
toplasmic rearrangements apparently do not occur in the
fertilized egg in appendicularian tunicates (Holland et a/.,
1988), which may be closest to the stem tunicate (Remane
etal.. 1976; Holland, 1991).
Acknowledgments
We are grateful to J. M. Lawrence of the University of
South Florida, Tampa. Florida, for the use of his labo-
ratory during the breeding season of B. floridae. This work
was supported in part by a National Science Foundation
grant No. NSF DCB 87-12888 to A. T. C. Carpenter,
Biology Department, University of California, San Diego,
who generously allowed LZH the use of her electron mi-
croscope.
Literature Cited
Aizenstadt, T. B., and N. S. Gavaeva. 1987. The perinuclear bodies
(nuage) in the developing germ cells of the lancelet Branchiostoma
laneeolatum. (In Russian). Tsitologia 29: 137-141.
Boveri, T. 1892. Liber die Bildungsstatte der Geschlechtdrusen und die
Entstehung der Genitalkamern beim Amphioxus. Anal. An:. 7: 1 70-
181.
Brien, P., and A. Dalcq. 1948. Generalites sur les cordes. Pp. 535-543
in Traiti- de Zoologie. Vol. 1 1 , P.-P. Grasse, ed. Masson et Cie, Paris.
Cerfontaine, P. 1906. Recherches sur le developpement de \'Amphioxit\
Arch. Biol. 22:229-418.
Conklin, E. G. 1905a. Organization and cell-lineage of the ascidian
egg. J. Acad. Nat. Sci. Philadelphia Ser. 2 13: 5-1 19 + plates I-XI.
Conklin, E. G. 1905b. Organ-forming substances in the eggs of ascidians.
Biol. Bull 8: 205-230 + plate XI.
Conklin, E. G. 1932. The embryology of amphioxus. J. Morphol 54:
69-151.
Conklin. E. G. 1933. The development of isolated and partially sepa-
rated blastomeres of amphioxus. / Exp. Zoo/. 64: 303-375.
Davidson, E. H. 1986. Gene Activity in Early Development. Third Edi-
tion. Academic Press. Orlando, FL. 670 pp.
Drach. P. 1948. Remarques sur les rapports des Cephalocordes et des
vertebres. Pp. 1031-1037 in Traite de Zoologie. Vol. 11. P.-P. Grasse,
ed. Masson et Cie, Paris.
Eddy, E. M. 1975. Germ plasm and the differentiation of the germ
cell line. Int. Rev. Cylol. 43: 229-280.
Erwin, D. II. 1991. Metazoan phylogeny and the Cambrian radiation.
77?E£6: 131-134.
Ghiselin, M. T., K. G. Field, G. J. Olsen, D. J. Lane, R. A. Raff, E. C.
Raff, and N. R. Pace. 1986. A phylogenetic tree of chordate subphyla
based on 1 8s ribcsomal RNA sequences. Am. Zoo/. 26: 92A.
Guraya, S. S. 1968. Cytochemistry of yolk elements in the amphioxus
egg. Z. Zellforsch. 79: 326-331.
Guraya, S. S. 1979. Recent advances in the morphology, cytochemistry
and function of Balbiani's vitelline body in animal oocyte. Int. Rev.
Cytol. 59: 24-32.
Guraya, S. S. 1983. Cephalochordata. Pp. 735-752 in Reproductive
Biology of Invertebrate*. Vol. 1, K. G. Adiyodi and R. G. Adiyodi.
eds. John Wiley and Sons. Chichester.
Hamaguchi, M. S., Y. Hamaguchi, and Y. Hiramoto. 1986. Micro-
injected polystyrene beads move along astral rays in sand dollar eggs.
Dev. Growth Differ. 28: 461-470.
Hatschek, B. 1882. Studien u'ber Entwicklung des Amphioxus. Arb.
Zool. Inst. Univ. Wien 4: 1-88 + plates I-IX.
Hatschek, B. 1893. The Amphioxus and its Development. (J. B. Tuckey,
translator). Swan, Sonnenschein and Co.. London. 181 pp.
Hay, B., L. Ackerman, S. Barbel, L. Y. Jan, and Y. N. Jan. 1988. Iden-
tification of a component of Drosophila polar granules. Development
103: 625-640.
Heggeness, M. H., M. Simon, and S. J. Singer. 1978. Association of
mitochondria with microtubules in cultured cells. Proc. Nail. Acad.
Sci. USA 75: 3863-3866.
Hirakow, R., and N. Kajita. 1990. An electron microscopic study of
the development of amphioxus, Branchiostoma belcheri tsingtauense:
cleavage. / Morphol. 203: 331-344.
Hirakow, R., and N. kajila. 1991. Electron microscopic study of the
development of amphioxus, Branchiostoma belcheri tsingtauense:
the gastrula. / Morphol. 207: 37-52.
Holland, L. Z. 1988. Spermatogenesis in the salps Thalia denwcratica
and Cydosalpa afjinis (Tunicata: Thaliacea): an electron microscopic
study. J. Morphol. 198: 189-204.
Holland, L. Z. 1992. The phylogenetic significance of tunicate sperm
morphology. Proc. Sixth Int. Congress Spermatology (in press).
Holland, L. Z., G. Gorsky, and R. Eenaux. 1988. Fertilization in Oi-
kopleura dioica (Tunicata, Appendicularia): acrosome reaction, cor-
tical reaction and sperm-egg fusion. Zoomorphologv 108: 229-243.
Holland, N. D., and L. Z. Holland. 1989a. Fine structural study of the
cortical reaction and formation of the egg coats in a lancelet (=am-
phioxus), Branchiostoma floridae (Phylum Chordata: Subphylum
Cephalochordata = Acrania). Biol. Bull. 176: 1 1 1-122.
Holland, N. D., and L. Z. Holland. 1989b. The fine structure of the
testis of a lancelet (=amphioxus), Branchiostoma floridae (Phylum
Chordata: Subphyulum Cephalochordata = Acrania). Acta Zool.
(Stockh.) 70: 2 1 1 -2 1 9.
Holland, N. D., and L. Z. Holland. 1991. The fine structure of the
growth stage oocytes of a lancelet (=amphioxus), Branchiostoma
laneeolatum. Invert. Reprod. Dev. 19: 107-122.
Jefferies, R. P. S. 1986. The Ancestry of the Vertebrates. British Mu-
seum (Natural History). London. 376 pp.
Jeffery, W. R. 1984. Pattern formation by ooplasmic segregation in
ascidian eggs. Biol. Bull. 166: 277-298.
Kobayakawa, Y. 1988. Role of mitotic asters in accumulation of pig-
ment granules around nuclei in early amphibian embryos. J. Exp.
Zool. 248: 232-237.
Kowalevsky, A. 1865. Istoriya razvitiya Amphioxus lanceolatus. (In
Russian, not seen by the present authors.) Magister's Thesis, St. Pe-
tersburg University, Russia.
Kowalevsky, A. 1867. Entwickelungsgeschichte des Amphioxus lan-
ceolatus. Mem. Acad. Imperial Sci. St. Petersbowg. Ser 711: 1-17
+ plates I-II1.
Lwoff, B. 1892. Ueber einige wichtige Punkte in der Entwicklung des
Amphioxus. Biol. Centralblail 12: 729-744.
96
L. Z. HOLLAND AND N. D. HOLLAND
Manes, M. E., and F. D. Barbieri. 1977. On the possibility of sperm
aster involvement in dorso-ventral polarization and pronuclear mi-
gration in the amphibian egg. J. Embryol. Exp. Morplwl. 40: 187-
197.
Nakazato, S., and K. Ikenishi. 1989. Monoclonal antibody production
against a subcellular fraction of vegetal pole cytoplasm containing
the germ plasm of Xenopus 2-cell eggs. Cell Differ. De\. 27: 163-
174.
Nelson, L., R. Lorentzon, L. Boquist, and S. Lovtrup. 1982. Mor-
phological differentiation of mitochondria in the early amphibian
embryo. Exp. Cell Res. 137: 25-29.
Nicander, L., B. A. Afzelius, and I. Sjoden. 1968. Fine structure and
early fertilization changes of the animal pole in eggs of the river
lamprey. Lampelra fluviatilis. J. Embryol. Exp. Morplwl. 19: 319-
326.
Nieuwkoop, P. D. 1991. The different origin of the primordial germ
cells (PGCs) in various groups of vertebrates. Proc. A' Ned. Akad.
Wet. 94: 103-110.
Nieuwkoop, P. D., and L. A. Sutasurya. 1976. Embryological evidence
fora possible polyphyletic origin of the recent amphibians. / Embryol.
Exp. Morphol. 35: 159-167.
Nieuwkoop, P. D., and L. A. Sutasurya. 1979. Primordial Germ Cells
in the Chordates Embryogenesis and Phylogenesis. Cambridge Uni-
versity Press. Cambridge, London. 187 pp.
Or ml. mi, G. 1958. Cleavage and development of egg fragments in as-
cidians. Ada Embryol Morplwl. Exp. 1: 247-272.
Phillips, C. R. 1982. The regional distribution of poly(A) and total
RNA concentrations during early Xenopus development. / Exp. Zoo/.
223: 265-275.
Phillips, C. R. 1985. Spatial changes in poly(A) concentrations during
early embryogenesis in Xenopus laevis: analysis by in situ hybridiza-
tion. Dev. Biol. 109: 299-310.
Remane, A., V. Storch, and U. Welsch. 1976. Systematische Zoologie.
Sliimme des Tierreichs. Fischer, Stuttgart. 678 pp.
Ressom, R. E., and K. E. Dixon. 1988. Relocation and reorganization
of germ plasm in Xenopus embryos after fertilization. Development
103: 507-518.
Sardet, C., J. Speksnijder, S. Inoue, and L. JafTe. 1989. Fertilization
and ooplasmic movements in the ascidian egg. Development 105:
237-249.
Sawada, T. 1988. The mechanism of ooplasmic segregation in the as-
cidian egg. Zoo/. Sci. 5: 667-675.
Sawada, T., and G. Schatten. 1981. Microtubules in ascidian eggs dunng
meiosis, fertilization, and mitosis. Cell Motil. Cytoskeleton 9: 219-
230.
Schatten, G., and H.Schatten. 1988. Effects of motility inhibitors dunng
sea urchin fertilization. Exp. Cell Res 135: 31 1-330.
Schliwa, M. 1984. Mechanisms of intracellular organelle transport.
Pp. 1-81 in Cell and Muscle Motility. Vol. 5, J. W. Shaw.ed. Plenum
Publ. Co., New York.
Shimizu, T. 1989. Asymmetric segregation and polarized redistribution
of pole plasm dunng early cleavages in the Tuhifex embryo: role of
actin networks and mitotic apparatus. Dev. Growth Differ. 31: 283-
297.
Sobotta, J. 1895. Die Befructung des Eies von Amphioxus lanceolatus.
Vorlaufige Mitteilung. Anal. An:. 11: 129-137.
Sobotta, J. 1897. Die Reifung und Befruchtung des Eies von Amphioxus
lanceolatus Arch. Mikr Anal. 50: 15-71 + plates II-V.
Speksnijder, J. E., L. F. JafTe, and C. Sardet. 1989. Polarity of sperm
entry in the ascidian egg. Dev Biol. 133: 180-184.
Talevi, R., and B. Dale. 1986. Electrical characteristics of ascidian egg
fragments. Exp Cell Res 162: 539-543.
Tung, T. C., S. C. Wu, and Y. Y. F. Tung. 1958. The development of
isolated blastomeres of amphioxus. Sci. Sin. 7: 1280-1320.
Tung, T. C., S. C. Wu, and Y. Y. F. Tung. 1960a. The developmental
potencies of the blastomere layers in amphioxus egg at the 32-cell
stage. Sci. Sin. 9: 119-141.
Tung, T. C., S. C. Wu, and Y. Y. F. Tung. 1960b. Rotation of the
animal blastomere in amphioxus egg at the 8-cell stage. Sci. Record.
N. S. 4: 389-394.
Tung, T. C., S. C. Wu, and Y. Y. F. Tung. 1962a. The presumptive
areas of the egg of amphioxus. Sci. Sin. 11: 629-645.
Tung, T. C., S. C. Wu, and Y. Y. F. Tung. 1962b. Expenmental studies
on the neural induction in amphioxus. Sci. Sin. 11: 805-820.
Vale. R. D., B. J. Schnapp, T. S. Reese, and M. P. Sheetz.
1985. Organelle, bead, and microtubule translocations promoted
by soluble factors from the squid giant axon. Cell 40: 559-569.
van der Stricht, O. 1896. La maturation et la fecondation de 1'oeuf
d' Amphioxus lanceolatus. Arch. Biol. 14: 469-495 + plates XX, XXI.
van VVijhe. 1893. Ueber Amphioxus Anal. An:. 8: 152-172.
Wakahara, M. 1989. Specification and establishment of dorsal-ventral
polarity in eggs and embryos of Xenopus laevis. Dev. Growth Differ.
31: 197-207.
Wall, R. 1990. This Side Up. Spatial Determination in the Early De-
velopment oj Animals. Developmental and Cell Biology Series 24, P.
W. Barlow, D. Bray, P. B. Green, J. M. W. Slack, eds. Cambridge
University Press. Cambridge. 436 pp.
Wickstead, J. H. 1975. Chordata: Acrania (Cephalochordata). Pp. 283-
3 1 9 in Reproduction in Marine Invertebrates, Vol. 2, A. C. Giese and
J. S. Pearse, eds. Academic Press, New York.
Willey, A. 1894. Amphioxus and the Ancestry of the Vertebrates. Mac-
millan and Co., New York. 316 pp.
Wilson, E. B. 1893. Amphioxus. and the mosaic theory of development.
J Morplwl. 8: 579-639.
Yamaguchi, Y., K. Murakami, M. Furusawa, and J. Miwa. 1983.
Germline-specific antigens identified by monoclonal antibodies in
the nematode Caenorhabditis elegans. Dev. Growth Differ. 25: 121-
131.
Yisraeli, J. K., S. Sokol, and D. A. Melton, 1989. The process of lo-
calizing a maternal messenger RNA in Xenopus oocytes. Development
1989(suppl.): 31-36.
Zalokar, M., and C. Sardet. 1984. Tracing of cell lineage in embryonic
development of Phallusia mammillata (Ascidia) by vital staining of
mitochondria. Dev Biol 102: 195-205.
Reference: Biol. Bull 182: 97-104. (February, 1992)
The Divergence of Species-Specific Abalone Sperm
Lysins is Promoted by Positive Darwinian Selection
YOUN-HO LEE AND VICTOR D. VACQUIER
Marine Biology Research Division 0202, Scripps Institution of Oceanography,
University of California, San Diego, La Jolla, California 92093
Abstract. Recognition by sperm lysin of the egg vitelline
envelope may be one component in determining the spe-
cies-specificity of fertilization in abalones. The amino acid
sequences of lysin proteins of seven California abalone
species were deduced from the cDNA sequences. This is
the first extensive comparison of a gamete recognition
protein from congeneric species. Each prelysin has a highly
conserved signal peptide of 1 8 amino acids, followed by
a mature sequence of 136-138 residues. Of 136 aligned
positions, 68 have the same amino acid in all seven se-
quences. The % identity relative to the red abalone lysin
sequence is: white 90%, flat 83%, pinto 82%, pink 78%,
black 71%, and green 65%. Hydropathy plots and a dis-
tance tree of the seven lysins show that red, white, and
flat lysins are more closely related to each other than to
the lysins of the other four species. A hypervariable, spe-
cies-specific, domain exists in all sequences between po-
sitions 2-12. Amino acid replacements between any two
lysins are mostly nonconservative. Analysis of the cDNA
sequences shows the number of nonsynonymous substi-
tutions (amino acid altering) exceeds the number of syn-
onymous substitutions (silent) in 20 of the 21 pairwise
comparisons of the seven sequences, indicating that pos-
itive Darwinian selection must promote the divergence
of lysin sequences.
Introduction
A striking feature of fertilization is the species specificity
of sperm-egg interaction in mammals (O'Rand, 1988;
Yanagimachi, 1988a, 1988b; Roldan and Yanagimachi,
1989) and invertebrates (Giudice, 1973; Summers and
Received 25 July 1 99 1 ; accepted 1 1 October 1991.
Abbreviations: Mr, relative molecular mass; VE, vitelline envelopes
of abalone eggs.
Hylander, 1975, 1976; Osanai and Kyozuka, 1982).
Sperm-egg mixtures from the same species usually yield
zygotes more efficiently than cross-species mixtures. Al-
though cross-species hybrid zygotes can be obtained in
mammals and invertebrates, the general observation is
that much higher concentrations of sperm are needed in
the insemination mixture to achieve fertilization. Blocks
to cross-species fertilization can occur at four points in
the process: induction of the sperm acrosome reaction by
components of the egg surface, adhesion of sperm to the
egg envelope, sperm penetration of the egg envelope, and
fusion of sperm and egg cell membranes. In echinoderms,
the greatest barrier to cross-species fertilization is the fail-
ure of sperm to adhere to the egg vitelline envelope (Sum-
mers and Hylander, 1975, 1976); in mammals it is the
failure of sperm to adhere to and penetrate the egg zona
pellucida (O'Rand, 1988; Yanagimachi, 1988a,b; Roldan
and Yanagimachi, 1989).
The divergence of gamete recognition proteins may be
important in the establishment of barriers to cross-fertil-
ization between populations. This may be an important
factor in the speciation of marine invertebrates using ex-
ternal fertilization. To learn how species-specific gamete
recognition proteins have diverged during evolution, we
have studied a protein from abalone sperm. Abalones are
marine archeogastropods of the genus Haliotis. Approx-
imately 70 extant species occur on coastlines of the world,
eight of them on the Pacific Coast of North America.
Although abalones are members of an ancient group of
gastropods, the genus Haliolis is relatively recent, most
fossils being from the Miocene (5-25 million years; Lind-
berg, 1991).
The abalone egg is contained within a glycoproteina-
ceous vitelline envelope (VE) about 0.6 /im in thickness
(Lewis et a/., 1982). The spermatozoon possesses a rela-
tively enormous acrosome granule (Lewis et ai, 1980;
97
98
Y.-H. LEE AND V. D. VACQUIER
Shiroya and Sakai, 1983) containing two abundant pro-
teins of about Mr 18,000 and 16,000 (Lewis et a/., 1982).
During fertilization, the sperm attaches to the egg VE,
the acrosome granule opens, and the two proteins are
secreted. A hole in the VE, 3 ^m in diameter, is created
in seconds, and the sperm passes through it to fuse with
the egg (Lewis et a/., 1982; Sakai et ai, 1982). Partial
amino acid sequence analysis (41 residues of the NH2-
terminus) shows that the Mr 1 8,000 protein is not a pre-
cursor of the Mr 16,000 protein (Vacquier, unpubl.).
When the purified Mr 16,000 acrosomal protein (sperm
lysin) is added to eggs, the VE rapidly dissolves by a non-
enzymatic mechanism; VE glycoproteins are not degraded
and no new NH:-termini are formed (Haino-Fukushima,
1974; Lewis et al., 1982;Hoshi, 1985). As previously dis-
cussed (Lewis et al., 1982; Hong and Vacquier, 1986;
Baginski et al., 1990; Vacquier et al., 1990), lysin may
act by competing for hydrogen and hydrophobic bonds
that hold the glycoproteinaceous fibers of the VE together.
The cDNA for pink and red abalone sperm lysins had
been previously cloned and sequenced (Vacquier et al.,
1990). To learn about the evolutionary divergence of
sperm lysin in California abalones, the polymerase chain
reaction was used to generate double stranded cDNA from
five additional species. The analysis of the seven deduced
amino acid sequences of lysin is the first extensive com-
parison of a gamete recognition protein in congeneric
species. We were surprised to find that the divergence of
the lysin sequences is promoted by positive Darwinian
selection.
Materials and Methods
The seven species of abalone used in this study were:
Haliotis corrugata (pink, M34389), H. cracherodii (black,
M5997 1 ), H.fulgens (green, M59972), H. kamtschatkana
(pinto, M59970), H. rufescens (red, M34388), H. sorenseni
(white, M59968), and H. walallensis (flat, M59969). The
GenBank cDNA sequence accession number follows the
common name of each species. The testes of male abalone
were removed and poly A+ RNA isolated as described
(Chomczynski and Sacchi, 1987; Vacquier et al., 1990).
Northern blot analysis with a full length red abalone lysin
cDNA as the probe, revealed a single band of hybridization
of approximately 660 nucleotides in all seven species
(Vacquier et ai, 1990; and unpubl.). Oligonucleotide
primers were synthesized to the 5' end of the previously
reported red and pink cDNA sequences (primer 6; GAA-
CAGATTACAAG^rGAAGCTGT; the italicized ATG
being the initiation codon), and to the complementary
strand of the 3' end of the sequence adjoining the poly A
tail (primer 7; TAGTA AATCTA TTTA 7TCTGGAAT,
the italicized being the complement of the poly A signal
sequence; Vacquier et al., 1990).
Two to ten ^g of poly A+ RNA were used for first
strand synthesis (Frohman, 1990). The RNA was washed
twice in 1 ml 80% ethanol, dried, and dissolved in 7 ^1
water containing 3 n\ of primer 7 (30 pMol). The tube
was heated to 95 °C for 5 min and then placed on ice for
10 min, followed by a 5-s centrifugation (a quick spin).
A reaction mixture of 10 /tl was added to the tube [the
10 ^1 contained: 2.5 /ul dNTP mix at 2 mM of each nu-
cleotide; 2.0 n\ 10 X RTC buffer (Frohman, 1990); 1.0 n\
human placenta! RNase inhibitor (Promega, Madison,
Wisconsin); 2.0 n\ MuLV reverse transcriptase (400 units);
and 2.5 j*l water]. This mixture of 20 /u' was incubated
for 1 h at 37°C. An additional 1 /tl of MuLV reverse
transcriptase was then added (200 units), and the incu-
bation continued 1 h at 45 °C. Following incubation, the
sample was diluted with 2 ml 0. 1 X TE ( 1 mM Tris, 0. 1
mM EDTA, pH 8.0) and concentrated to 50 ^1 with a
Centricon-30 microconcentrator (Amicon Inc., Beverly,
Massachusetts). The redilution in 2 ml 0. 1 X TE and con-
centration to 50-100 n\ was done three times. Sixteen ^1
of this first strand cDNA product was used for the second
strand synthesis.
To each tube of 16 /il was added 33 ^1 of a mixture of
5 n\ PCR buffer (Frohman, 1990), 2.5 (A dimethylsulf-
oxide, 5.0 n\ dNTPs at 2 mA/ each nucleotide, 6.0 jul
primer 7 in water (30 pMol), 6.0 ^1 primer 6 (30 pMol),
and 8.5 n\ water. The 49 n\ volume (in a 0.5 ml tube) was
heated to 95°C for 5 min and cooled slowly ( 1 h) to 50°C.
After a quick spin, 0.5 ^1 of Taq polymerase was added
(2.5 units, Perkin-Elmer-Cetus, Emeryville, California),
the tube vortexed gently, and 50 //I mineral oil added.
The tube was incubated for 15 min at 37°C followed by
40 min at 72°C. Amplification of the lysin cDNA was
accomplished in a temperature cycler by 40 cycles of 94°C
for 1 min, 45°C for 1 min, and 72°C for 2 min. An ad-
ditional 0.5 n\ of the Taq enzyme was added after the first
20 cycles. Following the last cycle, the temperature was
held at 72°C for 15 min. The tubes were cooled to 23°C
by a quick spin in a microfuge and 1 .0 ^\ of KJenow
fragment added, and the incubation continued for 30 min
at 23°C. Agarose gel electrophoresis of the 50 ^1 reaction
mixture showed the presence of one product of amplifi-
cation of approximately 650 nucleotides.
The amplified double stranded cDNA was purified ei-
ther by three cycles of dilution in 2 ml of 0.1 X TE and
concentration to 50-200 n\ with a Centricon-30, or by
separating the product by electrophoresis in 1% agarose
in 0.5 X TAB (0.04 M Tris acetate, 0.001 M EDTA. pH
8.0). The 650 base pair band was excised from the gel and
the cDNA purified with Prep-A-Gene (Biorad Labora-
tories, Richmond, California). ThecDNA was quantitated
by spectrophotometry and 1 ^g aliquots stored in 100%
ethanol at -20°C. For sequencing, 1 ,ug of DNA was
washed twice in 1 ml 80%' ethanol and dried. The DNA
DIVERGENCE OF ABALONE SPERM LYSINS
99
RED
WHITE
FLAT
PINTO
PINK
BLACK
GREEN
RED
WHITE
FLAT
PINTO
PINK
BLACK
GREEN
RED
WHITE
FLAT
PINTO
PINK
BLACK
GREEN
RED
WHITE
FLAT
PINTO
PINK
BLACK
GREEN
-18 -1 1 21
MKLLVLCIFAMMATLAMSR-SWHYVEPKFLNKAFEVALKV
. - R
. N F
. T
p
F
V
. . . F . . .
. T .
• Q
. . E
L . . .
V .
. H R
FRF
I P
H
Y
T
R E
V ...
V .
D
Y Q F
• Q
H
n
Y
T
R . . .
. W
V .
. V .
. - R
. T F
. R
Y
H
Y
T
. . Y .
. T M . I
22 40 61
QIIAGFDRGLVKWLRVHGRTLSTVQKKALYFVNRRYMQTH
. T
. A . . . S
. . H .
E .
. V R W .
. K
. A . . .
. . . N . G P
. . H .
E .
. T
. . . . G R
E
. T
. N . . . G
. N E N
R V
. S .
. Q
. T A .
. R
. T N N
. T ,
F .
62 80 101
WANYMLWINKKIDALGRTPVVGDYTRLGAEIGRRIDMAYF
. T . . . . D .
V
T .
.Q VR.T.
.Q V..TN
.Q VKR. .K
D .
V F
. A . . A
R
P...A..S
P...A..RA G.
PAA . V . . V F
102 120 136
YDFLKDKNMIPKYLPYMEEINRMRPADVPVKYM--GK
I R
R - -
N...GR S..I
N...GR S..I
K..SGR M....Q A.. I --
N..NGR A N R - -
N. . .NRV RRL.N. . .E. . . .ANRNP
N. .SGRK. . . P . S A . .AKL.AL N H - -
Figure 1. Ammo acid sequences of the seven abalone sperm lysins. Dots denote identity to the red lysin
sequence, dashes are for alignment, and numbering refers to the red lysin sequence. The signal peptide spans
positions -18 to -1. The single letter amino acid code is used.
pellet was dissolved in 10 ^1 of sequenase buffer (U. S.
Biochemicals, Cleveland, Ohio) containing 10 pMol of a
sequencing primer. Eight different oligonucleotide primers
were used for sequencing, all of them synthesized so as
to correspond to the red abalone lysin cDNA sequence
(Vacquier et ai, 1990). The tube containing the mixture
of DNA and primer was heated to 95°C for 5 min and
snap frozen 5 min in a dry ice ethanol bath, then placed
in an aluminum block precooled to -20°C, which was
allowed to warm to 23 °C over a 2-h period. Following a
quick spin, the Sequenase protocol was performed and
the sequences of both strands of cDNA determined a
minimum of two times. The cDNA and amino acid se-
quences were computer aligned and listed in order of sim-
ilarity using the progressive alignment and tree building
program given in Feng and Doolittle (1990). Hydropathy
plots with a window of seven amino acids were done by
the method of Kyte and Doolittle (1982).
To determine whether amino acid replacement between
any two lysin sequences conserved the class of residue,
the 20 amino acids were divided into 5 classes following
the structural considerations of Dickerson and Geis
(1983). Synonymous and nonsynonymous nucleotide
substitutions were computed by the methods of Li et al.
(1985) and Nei and Gojobori (1986).
Results
Deduced amino acid sequences
The deduced amino acid sequences, aligned and listed
in order of descending similarity (Feng and Doolittle,
1990), are presented in the single letter code in Figure 1.
Assignments of the initiation methionine (M at position
-18), the signal sequence of 18 amino acids, the NH2-
terminal residue of the mature lysins being arginine (R
at position 1 ), and the COOH-terminal residue being ly-
sine (K at position 136) have been previously presented
(Vacquier et ai, 1990). In Figure 1, dots denote identity
to the red abalone lysin sequence and dashes are for align-
ment. The signal sequences (positions —18 to —1) have
been highly conserved during evolution and are typical
of eukaryotes (von Heijne, 1985). Neither cysteines nor
100
Y.-H. LEE AND V. D. VACQUIER
sites for N-linked glycosylation are found in the mature
lysins. The mature pink abalone lysin is 137 residues,
black abalone lysin is 138, and the lysins of the five other
species are 1 36 residues in length. Of 1 36 aligned positions,
68 (50%) have the same amino acid in all seven species.
The two longest regions of perfect identity are the eight
residues between positions 88-95 and the 1 1 residues be-
tween positions 52-62. There are four occurrences of two
contiguous positively charged amino acids in all seven
sequences (positions 47-48, 55-56, 71-72, and 94-95).
The percent identity in amino acid residues for the 21
pairwise comparisons of the seven lysin sequences (Table
I) shows the decrease in similarity progressing from red
lysin to green lysin. Green abalone lysin is equally dissim-
ilar from each of the other six lysins, the percent identity
varying from 63 to 65%. The region of greatest difference
among all seven sequences is the 1 1 residue segment com-
prising positions 2-12 (Fig. 1 ). In this region, no two spe-
cies have the identical sequence. Considerable difference
in charge distribution is seen in this hypervariable seg-
ment. For example, the red, flat, and pinto lysins have a
net charge of + 1 , whereas pink abalone lysin has a net
charge of +6.
Hydropathy plots and branching order
Hydropathy plots of the seven mature lysin sequences
are of value in showing subtle differences throughout the
sequences (Fig. 2). The plots of the hypervariable domain
of positions 2-12 (shaded) are in most cases species-spe-
cific. The upper three plots (red, white, flat) are quite sim-
ilar, all having a large hydrophobic domain between res-
idues 1 5-30. The pinto is clearly different from the top
three, this large hydrophobic domain being reduced and
followed by a hydrophilic domain centered at position
30. The pink lysin has a hydrophilic domain centered at
residue 60 that is more similar in shape to the one in the
black and green lysins than it is to the other four species.
There is a moderately hydrophobic domain at about po-
sition 70 in pink lysin shared with only the black species.
The pinto and green are the only two lysins having a large
Table I
Percent identily of amino acid residues in 21 pairwise comparisons
in 136 aligned positions of seven lysins
Species Red
White
Flat
Pinto
Pink
Black
White
Flat
90
83
88
Pinto
82
85
76
Pink
78
80
77
72
Black
71
72
72
65
78
Green
65
64
65
63
65
65
JU
POSITION
Figure 2. Hydropathy plots of the seven abalone lysin sequences
using a window of seven residues. Hydrophobic values are positive and
hydrophilic values negative. The hypervanable domain of 2-12, and the
two invariant domains of 52-62 and 88-95 are shaded. R, red; W, white;
F, flat; Pt, pinto; Pk, pink; B, black; and G, green abalone lysins.
hydrophobic peak close to position 100. The plots for the
black and green are clearly distinct from the other species
in having two large hydrophilic domains centered about
positions 45 and 60. However, the plot of the black lysin
shows a large hydrophilic domain around position 125
which is not present in the green. The regions of 1 1 (po-
sitions 52-62) and eight (positions 88-95) amino acids
that are invariant in all seven sequences are shown as
shaded zones (Fig. 2). Both these regions are amphipathic,
being hydrophobic in the NHrterminal direction and hy-
drophilic in the COOH-terminal direction.
A distance tree depicting the branching order of the
seven lysin sequences (Fig. 3) shows red and white lysins
to be the most closely related proteins. The black and
green lysin sequences are the most divergent; they are far
from the other five sequences and also far from each other.
Amino acid replacements and nucleotide substitutions
The 2 1 possible pairwise comparisons of the 7 lysin
sequences were analyzed to determine the fraction of
DIVERGENCE OF ABALONE SPERM LYSINS
101
Table II
The replacement ofamino acids in mature lysins is nonconservative
Species
Red White Flat Pinto
Pink
Black
White 7,,
Flat n/23
10;
/1 6
Pinto 5/25
/2I
'%3
Pink %,
%6
'2Ao '/37
Black '%o
As
11; 10;
As /«
5/27
Green 18/4g
"A,
21/ I9/
ki Ai
16; 15;
/46 /48
amino acid replacements that were between residues of
the same amino acid class (conservative replacement). The
data are shown in Table II, where the numerator is the
number of replacements between amino acids of the same
class, and the denominator is the total number of replace-
ments in each pairwise comparison. In all but two com-
parisons (flat X red and flat X white), conservative amino
acid replacements are far below 50%. In summary, the
majority of amino acid replacements between any two
lysins involves switching the class of residue.
The number of synonymous (Ds) and nonsynonymous
(Dn) nucleotide substitutions per site were computed for
the 2 1 pairwise comparisons of the seven lysin cDNA
sequences (Nei and Gojobori, 1986). The data (Table III)
show that in all but one comparison (flat X green) Dn is
greater than Ds. In 6 of the 2 1 comparisons, the difference
is significant at the 5% level, and in two at the 0.5% level.
These data show that positive Darwinian selection is pro-
moting the divergence of lysin sequences. Also, the closely
related sequences (Fig. 3) of red, white, flat, and pinto
abalone lysins exhibit the positive selection phenomenon
more strongly than do the more widely divergent se-
quences.
Discussion
Amino acid sequences
Homology among the seven mature lysins is readily
apparent (Fig. 1 ). The sequences align perfectly in 952 of
955 amino acids. Lysins are constrained in length, varying
from 136 to 138 amino acids. There is conservation of
primary structure in that 68 of the 136 aligned positions
have the identical amino acid in all seven sequences. With
the exception of the hypervariable domain of positions
2-12, these invariant 68 positions are spread throughout
the lysin molecule with a slight concentration toward the
central portion of the sequence. Of the 68 invariant po-
sitions, 14 are occupied by residues that are highly con-
served (Graur, 1985; single letter code, W = 3, G = 5,
and Y = 6), and 24 by the group of seven amino acids
that are replaced most frequently in mammalian proteins
(Graur, 1985; T = 1, H = 1, Q = 2, F = 3, 1 = 4, M = 5,
Table HI
Percent synonymous (Ds) and non-synonymous (Dn) nuc/eolide substitutions per site
Species
Ds
(SE)
Dn
(SE)
Dn
Ds
d =
Dn - Ds
SE
Red
x White
1.62
(1.32)
5.79
(1.40)
3.57
4.17*
1.92
Flat
2.61
(1.69)
10.08
(1.89)
3.86
7.47"
2.54
Pinto
2.76
(1.75)
10.71
(1.95)
3.88
7.95**
2.62
Pink
10.59
(3.53)
14.63
(2.33)
1.38
4.04
4.23
Black
11.08
(3.64)
21.76
(2.95)
1.96
10.68*
4.69
Green
21.27
(5.31)
24.93
(3.22)
1.17
3.66
6.21
White
X Flat
4.25
(2.16)
6.36
(1.47)
1.50
2.11
2.61
Pinto
3.30
(1.91)
8.39
(1.70)
2.54
5.09*
2.56
Pink
9.92
(3.40)
13.69
(2.25)
1.38
3.77
4.08
Black
11.42
(3.69)
20.82
(2.87)
1.82
9.40*
4.67
Green
21.12
(5.27)
25.44
(3.26)
1.20
4.32
6.20
Rat
X Pinto
5.45
(2.48)
13.59
(2.23)
2.49
8.14*
3.34
Pink
11.67
(3.71)
15.86
(2.45)
1.36
4.19
4.45
Black
13.42
(4.03)
22.70
(3.04)
1.69
9.28
5.05
Green
25.02
(5.84)
23.78
(3.13)
0.95
-1.24
6.63
Pinto
x Pink
9.76
(3.39)
18.05
(2.63)
1.85
8.29
4.29
Black
13.71
(4.12)
26.06
(3.31)
.90
12.35*
5.28
Green
22.66
(5.54)
24.49
(3.18)
.08
1.83
6.39
Pink
X Black
9.31
(3.28)
13.12
(2.19)
.41
3.81
3.94
Green
16.03
(4.45)
24.02
(3.15)
.50
7.99
5.45
Black
X Green
18.14
(4.80)
27.80
(3.46)
.53
9.66
5.92
* Significant at 5% level, ** at 0.5% level.
102
Y.-H. LEE AND V. D. VACQUIER
and L = 8). Because the occupancy of these 68 positions
is identical in all 7 lysins, we conclude that they are crucial
to lysin's role in fertilization in California abalones.
The hypervariable domain (positions 2-12) is strikingly
similar to the ligand binding domain of annexin II (Becker
el al, 1990), a membrane and lipid binding protein. An-
nexin II possesses an NH:-terminal 1 2 amino acid segment
(Ac-STVHEILCKLSL) that binds its ligand (pi 1 ). Ligand
binding induces the 12 residues to form a positively
charged amphipathic «-helix that becomes buried in p 1 1 .
The important structural features for binding between
annexin II and pi 1 are the hydrophilic residue in position
1 and the hydrophobic side chains at positions 3, 6, 7,
and 10. White abalone lysin has residues with hydrophobic
side chains at positions 3, 6, 7, and 10, and the same
positively charged residues at positions 4 and 9 (His and
Lys) as has annexin II. In red, pink, and pinto lysins, 3
out of 4 residues at positions 3, 6, 7, and 10 have hydro-
phobic side chains. Three of the seven lysins are positively
charged at position 4 and six out of seven at position 9.
The binding of lysin to its unknown VE ligand may thus
be similar to the binding of annexin II to pi 1 (Becker el
al., 1990).
Much has been learned about protein-protein recog-
nition by X-ray crystallographic studies of the binding of
proteases with their inhibitor proteins, and the binding
of antibody to antigen (Janin and Chothia, 1990). In the
protease-inhibitor complexes, 10-15 residues of the in-
hibitors make contact with 1 7-29 residues of the proteases.
These numbers are consistent with the size of the lysin
hypervariable domain. In antibodies, the antigen binding
sites are disproportionately rich in residues with aromatic
side chains. In the seven lysins, between positions 2-12,
26 residues of a total of 77 (34%) in all 7 sequences have
aromatic side chains, whereas by total amino acid com-
position, only 17% of lysin residues are aromatic. This
adds support to the concept that positions 2-12 in lysin
may be involved in the binding of its VE ligand. We have
not as yet quantitatively determined the ability of lysin
to dissolve egg VEs in all 2 1 pairwise combinations of the
7 species. However, we have demonstrated species spec-
ificity in the cross combinations of red and pink abalone
lysins and egg VEs proteins (Vacquier el al., 1990).
Positive Darwinian selection in lysin divergence
In most cases, when two orthologous proteins are di-
verging, the frequency of synonymous (silent) nucleotide
substitution (Ds) will be greater than that of nonsynon-
ymous (amino acid altering) substitution (Dn). If positive
Darwinian selection is promoting divergence of two pro-
teins, the converse will be true. In positive selection there
is adaptive value to alter the amino acid sequence. Positive
selection has been proven at the molecular level in the
following cases: the class I (Hughes and Nei, 1988; Hughes
el al., 1990) and class II (Hughes and Nei, 1989) major
histocompatibility complex antigens; the VH genes of im-
munoglobulins (Tanaka and Nei, 1989); the circumspo-
rozoite antigen in Plasmodium (Hughes, 1991); human
influenza A virus (Fitch et al., 1991); and the Adh locus
in Drosophila (McDonald and Kreitman, 1991). The
nonconservative nature of amino acid replacements be-
tween lysins (Table II) provided the clue that positive se-
lection might be promoting lysin divergence. Analysis of
lysin cDNA sequences by the method of Nei and Gojobori
(1986; Table III) shows that Dn exceeds Ds in 20 of 21
pairwise comparisons. Among the closely related se-
quences of the red, white, flat, and pinto abalone lysins
(Fig. 3), Dn shows statistically significant higher values
than Ds. Analysis by a similar method (Li et al., 1985)
yielded almost the same results. For example, in com-
parisons of red, white, flat, and pinto lysin sequences, the
average nonsynonymous value was 2.4 times greater than
the synonymous value. The most extreme comparison
was between the red and flat sequences, where the non-
synonymous value was 3.4 times greater than the syn-
onymous value. These data indicate a strong selective ad-
vantage in altering the amino acid sequence of lysin. This
is the first example of positive selection acting on a gamete
recognition protein. With the exception of the Adh locus
in Drosophila, the common attribute abalone lysins share
RED
WHITE
FLAT
-PINTO
PINK
BLACK
11.97
GREEN
Figure 3. Distance tree showing the branching relationships of the
lysin sequences. The root of the tree is placed arbitrarily at the midpoint.
The numbers on the branches represent relative evolutionary distances
(Feng and Doolittle. 1990).
DIVERGENCE OF ABALONE SPERM LYSINS
103
with the published examples of positive selection is the
involvement in extracellular recognition.
We cannot speculate about what could provide the se-
lective pressure acting on lysin divergence. The demon-
stration of positive selection does not prove that it is a
causative factor responsible for speciation in abalones.
Experimental evidence exists that abalone embryos tend
to settle near their parents (Prince et al, 1987), and that
genetic structure can occur within an abalone species in
two populations separated by three km (Brown, 1991).
Thus, speciation by geographic isolation probably occurs
in abalones. Although the demonstration of positive se-
lection in lysin divergence does not indicate how abalone
populations split into distinct species, the possibility that
it may accompany the speciation process should be con-
sidered. The statistically significant data showing positive
selection (Table III) are between the closely related abalone
species (Fig. 3). This suggests that a high frequency of
nonsynonymous substitution (Dn) accompanies initial
divergence, but Dn decreases as divergence increases. A
similar situation occurs with the class I major histocom-
patibility genes in which Dn is greater in intralocus as
compared to interlocus comparisons (Hughes and Nei,
1988). Thus, the reason that the positive selection data
set is so robust for lysin may be due to the relatively recent
appearance of these closely related species in the fossil
record (20 million years ago; Lindberg, 1991).
One might speculate that positive selection may cause
allelic variation in abalone sperm lysins. However, two
male pink abalones from San Diego, California and six
male red abalones (two from San Nicholas Island, two
from Mendocino, California, and two from San Diego)
yielded identical species-specific cDNA sequences in both
the 462 nucleotide open reading frame and in about 1 50
nucleotides of the 3' untranslated region containing the
poly A signal sequence (Vacquier et al., 1990). San Diego
and Mendocino are separated by roughly 800 km of
coastline, and by Point Conception, an important eco-
logical barrier to larval transport. We tentatively conclude
from these limited numbers of individuals that there is
no major allelic variation in lysin sequences in the red
abalone. The species-specific lysin sequences may thus be
well fixed in the extant California species. As pointed out
by a reviewer, there are currently no models to explain
these data; they represent a genuine mystery for future
research to solve.
In the class I major histocompatibility antigens, the
antigen recognition site exhibits positive selection, but the
different alleles share many conserved structural features
making their homology obvious over tens of millions of
years of evolution (Hughes and Nei, 1988). Abalone sperm
lysins are similar in that strong homology exists among
all seven lysins, yet 50% of the positions have species-
specific amino acid replacements. Knowing the sequences
of these seven sperm lysins begs the question as to the
nature of sequence variation in the VE ligands that are
the lysin "receptors" of the egg surface. We predict that
these ligands will show the same pattern of variation; that
is, some regions will be conserved in all species, while
others will be hypervariable and species-specific.
Acknowledgments
We thank J. W. Swanson, R. McConnaughey, Dr. D. L.
Leighton, and Dr. J. R. Pawlik for abalones. Dr. Wen-
Hsiung Li for his computer program, Drs. Masatoshi Nei
and Tatsuya Ota for their interest and considerable effort
in preparing the data presented as Table III. Discussions
with Drs. D. W. Smith, D.-F. Feng, R. F. Doo-
little, W. M. Fitch, J. E. Minor, R. J. Britten, S. R. Palumbi,
D. A. Powers, and F. Azam are gratefully acknowledged.
Supported by NIH Grant HD 12986 to V.D.V. and by a
Korean Government Overseas Scholarship to Y.-H.L.
This paper is dedicated to the memory of Professor Alberto
Monroy.
Literature Cited
Baginsky, M. L., C. D. Stout, and V. D. Vacquier. 1990. Diffraction
quality crystals of lysin from spermatozoa of the red abalone (Haliotis
rufescens). J. Biol. Cheni. 265: 4958-4961.
Becker, T., K. Weber, and N. Johnson. 1990. Protein-protein recog-
nition via short amphiphilic helices; a mutational analysis of the
binding site of annexin II for pi 1. EMBO J. 9: 4207-4213.
Brown, L. D. 1991. Genetic variation and population structure in the
blackhp abalone, Haliotis rubra. Aust. J. Mar. Freshwater Res. 42:
77-90.
Chomczynski, P., and N. Sacchi. 1987. Single-step method of RNA
isolation by acid guanidinium thiocyanate-phenol-chloroform ex-
traction. Anal. Biochem 162: 156-159.
Dickerson, R. E., and I. Geis. 1983. Hemoglobin, structure. Junction,
evolution and pathology Benjamin/Cummings Inc. Menlo Park, CA.
Feng, D.-F., and R. F. Doolittle. 1990. Progressive alignment and phy-
logenetic tree construction of protein sequences. Meth. Enzymol 183:
375-387.
Fitch, W. M., J. M. E. Leiter, X. Li, and P. Palese. 1991. Positive
Darwinian evolution in human influenza A viruses. Proc. Nail. Acad.
Sci. USA 88: 4270-4274.
Frohman, M. A. 1990. RACE: rapid amplification of cDNA ends. Pp.
28-38 in PCR Protocols, M. A. Innis, D. H. Gelfand, J. J. Sninsky
and T. J. White, eds. Academic Press, San Diego.
Giudice, G. 1973. Pp. 162-174 in Developmental Biology' of the Sea
Urchin Embryo. Academic Press. San Diego.
Graur, D. 1985. Amino acid composition and the evolutionary rates
of protein-coding genes. / Mol. Evol 22: 53-62.
Haino-Fukushima, K. 1974. Studies on the egg membrane lysin of 7V-
giila pfeifferi: the reaction mechanism of the egg membrane lysin.
Biochim. Biophys. Act a 352: 179-191.
Hong, K., and V. D. Vacquier 1986. Fusion of liposomes induced by
a cationic protein from the acrosome granule of abalone spermatozoa.
Biochemistry 25: 543-550.
Hoshi, M. 1985. Sperm lysins. Pp. 43 1 -462 in Biology of Fertilization,
C. B. Metz and A. Monroy, eds. Academic Press, San Diego.
Hughes, A. L. 1991. Circumsporozoite protein genes of malaria par-
asites (Plasmodium spp.): evidence for positive selection on immu-
nogenic regions. Genetics 127: 345-353.
104
Y.-H. LEE AND V. D. VACQUIER
Hughes, A. L., and M. Nei. 1988. Pattern of nucleotide substitution
at major histocompatibility complex loci reveals overdominant se-
lection. Nature 335: 167-170.
Hughes, A. L., and M. Nei. 1989. Nucleotide substitution at major
histocompatibility complex class II loci: evidence for overdominant
selection. Proc. Nail Acad. Sci. USA 86: 958-962.
Hughes, A. L., T. Ota, and M. Nei. 1990. Positive Darwinian selection
promotes charge profile diversity in the antigen binding cleft of class
I MHC molecules. Mat. Biol. Evol. 7: 515-524.
Janin, J., and C. Chothia. 1990. The structure of protein-protein rec-
ognition sites. J. Biol. Chem. 265: 16027-16030.
Kyte, J., and R. F. Doolittle. 1982. A simple method for displaying
the hydropathic character of a protein. J. Mot. Biol. 157: 105-132.
Lewis, C. A., D. L. Leighton, and V. D. Vacquier. 1980. Morphology
of abalone spermatozoa before and after the acrosome reaction. J.
Ultrastnict. Res. 72: 39-47.
Lewis, C. A., C. F. Talbot, and V. D. Vacquier. 1982. A protein from
abalone sperm dissolves the egg vitelline layer by a nonenzymatic
mechanism. Dev. Biol. 92: 227-240.
Li, W.-H., C.-I. Wu, and C.-C. Luo. 1985. A new method for estimating
synonymous and nonsynonymous rates of nucleotide substitution
considering the relative likelihood of nucleotide and codon changes.
AM Biol. Evol. 2: 150-174.
Lindberg, D. R. 1991. Evolution, distribution and systematics of Hal-
lotidae. In Abalone of the World: Biology. Fisheries and Culture, A.
Shepherd, M. J. Tegner, and S. A. Guzman Del Proo, eds. Blackwells
Scientific Publishers, London.
McDonald, J. H., and M. Kreitman. 1991. Adaptive protein evolution
at the Adh locus in Drosophila. Nature 351: 652-654.
Nei, M., and T. Gojobori. 1986. Simple methods for estimating the
numbers of synonymous and nonsynonomous nucleotide substitu-
tions. Mol. Biol. Evol. 3: 418-426.
O'Rand, M. G. 1988. Sperm-egg recognition and barriers to interspecies
fertilization. Gamete Res. 19: 315-328.
Osanai, K., and K. Kyozuka. 1982. Cross fertilization between sea urchin
eggs and oyster spermatozoa. Gamete Res 5: 49-60.
Prince, J. D., T. L. Sellers, W. B. Ford, and S. R. Talbot. 1987. Ex-
perimental evidence for limited dispersal of haliotid larvae (genus
Haliotis: Mollusca: Gastropoda). J. Exp. Mar. Biol. Ecol. 106:
243-263.
Roldan, E. R. S., and R. Yanagimachi. 1989. Cross-fertilization between
Syrian and Chinese hamsters. /. Exp. Zool. 250: 321-328.
Sakai, Y. T., Y. Shiroya, and K. Haino-Fukushirna. 1982. Fine struc-
tural changes in the acrosome reaction of the Japanese abalone, Hal-
iotis discus. Dev. Growth Differ. 24: 531-542.
Shiroya, Y., and Y. T. Sakai. 1983. Fine structure of the spermatozoon
in the Japanese abalone, Haliotis discus. J Wayo Women's Univ.
24: 253-267.
Summers, R. G., and B. L. Hylander. 1975. Species-specificity of ac-
rosomal reaction and primary gamete binding in echinoids. Exp.
Cell Res 96: 63-68.
Summers, R. G., and B. L. Hylander. 1976. Primary gamete binding:
the species-exclusive event of echinoid fertilization. Exp. Cell Res
100: 190-194.
Tanaka, T., and M. Nei. 1989. Positive Darwinian selection observed
at the variable region genes of immunoglobins. Mol. Biol. Evol. 6:
447-459.
Vacquier, V. D., K. R. Garner, and C. D. Stout. 1990. Species-specific
sequences of abalone lysin, the sperm protein that creates a hole in
the egg envelope. Proc. Natl. Acad. Sci. USA 87: 5792-5796.
von Heijne, G. 1985. Signal sequences: the limits of variation. J. Mol.
Biol 184:99-105.
Yanagimachi, R. 1988a. Mammalian fertilization. Pp. 135-185 in The
Physiology of Reproduction, E. Knobil and J. Neill, eds. Raven Press,
New York.
Yanagimachi, R. 1988b. Sperm-egg fusion. Curr. Top. Membr. Trans-
port 32: 3-43.
Reference: Biol. Bull. 182: 105-108. (February, 1992)
Embryos of Homarus americanus are Protected
by Epibiotic Bacteria
M. SOFIA GIL-TURNES AND WILLIAM FENICAL
Scripps Institution oj Oceanography, University of California, San Diego,
La Mia. California 92093-0236
Abstract. Embryos of the American lobster, Homarus
americanus, are remarkably resistant to infection by the
fungus Lagenidium callinectes, a pathogen of many crus-
taceans. The surfaces of healthy lobster embryos are cov-
ered almost exclusively by a single, Gram-negative bac-
terium, which grows in a dense mosaic pattern. In culture,
this bacterium produces a compound that completely in-
hibits the growth of the pathogenic fungus in vitro at 10
meg/ml. Large-scale fermentation, extraction, and sub-
sequent chromatographic purification led to the identi-
fication of the antifungal substance as 4-hydroxyphenethyl
alcohol (tyrosol), an antibiotic substance known to be
produced by terrestrial fungi.
Introduction
Like several other decapod crustaceans, the American
lobster Homarus americanus incubates its embryos ex-
ternally, and each female carries a large cluster comprising
up to 60,000 embryos. The embryos are attached to spe-
cialized abdominal pleopods until hatching some nine
months after fertilization (Cobb and Wang, 1985).
Throughout this long brooding period, during which the
female is said to be "in berry," the embryos are contin-
uously exposed to water-borne microorganisms. It is re-
markable that the seemingly unprotected embryos can
survive microbial encroachment. The phycomycetous
fungus Lagenidium callinectes is a pathogen of many
crustaceans. Larvae and juveniles of the American lobster,
when kept in unnatural conditions, e.g.. in aquaria, are
extremely vulnerable to infection by this fungus (Fisher
el al, 1976; Nilson el ai, 1976; Provenzano, 1985). In
contrast, Homarus embryos, even when detached from
the female, appear to be remarkably resistant to fungal
Received 31 July 1991; accepted 7 October 1991.
attack. This situation is highly analogous to that recently
observed for the estuarine shrimp Palaemon macrodac-
tylus (Gil-Turnes el al., 1989). Embryos of P. macwdac-
tylus were also impervious to attack by L. callinectes. In
that study, Palaemon embryos were found to host an epi-
biotic bacterium that produced 2,3-indolenedione, a
molecule toxic to L. callinectes. The intent of this research
was to compare Homarus americanus with Palaemon
macrodactylus, and to determine whether lobster embryos
are also protected by an association with symbiotic bac-
teria.
Materials and Methods
Embryo infection experiment with Lagenidium
callinectes
Gravid female Homarus americanus were collected in
the vicinity of Martha's Vineyard, Massachusetts, air
shipped to California, and maintained in aquaria at the
Bodega Marine Laboratory in Bodega Bay, California
(UC-Davis). The embryos were observed to be at different
stages of development. Ten groups of five embryo clusters
each were detached from the females. Each cluster was
rinsed with three aliquots of sterilized seawater and sub-
sequently suspended from a cotton thread in aerated 125
ml Erlenmeyer flasks containing 75 ml of sterile seawater.
A liquid suspension of Lagenidium callinectes was pre-
pared by homogenizing a 0.5 cm diameter agar core of
fungal hyphae in 10 ml of 2216 Difco Marine Broth me-
dium. The fungus grew for one week, and the culture was
then shaken vigorously to break up the hyphae. Aliquots
of 1 ml of this thick suspension were added to each ex-
perimental flask. Addition of fungal culture was repeated
after the fifth day on a daily basis for a period of two
weeks.
105
106
M. S. GIL-TURNES AND W. FENICAL
Isolation and culture of associated bacteria
In a typical experiment, five embryos from each animal
were homogenized in an autoclaved tissue grinder with
10 ml of sterile senvater. One drop of the homogenate,
and of 1/10 and i /1 00 dilutions, were plated on Difco
2216 Marine Agar plates. Colonies were removed and
subcultured after 1-3 weeks. Although three or four mor-
phologically variable colonies were generally observed, one
distinct bacterium (SGT-76), a salmon-colored, slow-
growing (at 21°C), Gram-negative rod, was consistently
obtained. In liquid culture, this strain inhibited the growth
of L. callinecles. Antifungal testing was performed by cut-
ting agar cores, 0.5 cm in diameter, from lawns of the
pure bacterium and placing them on agar plates approx-
imately 1 cm from agar cores containing radiating hyphae
of L. callinectes. Because of the inhibition observed, this
bacterium was selected for subsequent chemical studies.
Extraction and purification of the antifungal compound
The Gram-negative, salmon-colored bacterium, SGT-
76, was cultured, at 2 1 °C, in a 1 6-1 carboy using a medium
composed of 3 g BactoPeptone (Difco) and 5 g yeast ex-
tract per liter of seawater. The culture grew, with aeration,
for three weeks. The final pH of the medium was 8.8. The
entire culture was extracted twice with 4 1 ethyl acetate.
After evaporation of the combined solvents, the remaining
crude extract was fractionated by silica-gel vacuum flash
chromatography using variable amounts of ethyl acetate
in isooctane. The antifungal activity of each fraction was
determined by placing 0.5 mg of each dry fraction onto
a 0.5 cm paper disk and placing the disk at the edge of
fungal growth. The active compound eluted with 80%
ethyl acetate/isooctane. Final purification of the antifungal
compound was achieved by size exclusion chromatogra-
phy on Sephadex LH20 using a mixture of hexane/meth-
ylene chloride/methanol (2:5: 1 ). The purified compound
was characterized by infrared spectroscopy (IR), by high-
resolution mass (HRMS) and by 'H and I3C nuclear mag-
netic resonance spectrometry (NMR).
Scanning electron microscopy
Embryos were fixed in 2.5% glutaraldehyde in 3% saline
solution for a linimum of 24 h. After three rinses in 3%
saline solutio: >in, each specimen was transferred
to a solution of i ii um tetroxide in 3% saline solution
for nine minutes. were then stored in saline
solution overnight .equently dehydrated using an
acetone/distilled water •; ic-nce: 35% for 15 min, 50%
for 15 min, 75% for 30 min, 95% for 1 h, and 100% ethanol
for 12 h. Critical point drying was done under CO2 and
the gold coating thickness was 300 A. Electron micro-
graphs were obtained with a Hitachi Model 539 SEM.
Pure bacterial films were prepared as follows: a drop of a
3-day liquid culture was deposited on a small millipore
filter (0.25 n pore size) placed on an agar plate. As soon
as growth was visible, the specimens were fixed in 3%
formaldehyde and 3% glutaraldehyde in 0.2 M sodium
cacodylate trihydrate buffer (pH 7.4) for 1 h and then
washed three times for 5 min in 0.2 M cacodylate solution.
The specimens were then transferred to 2% osmium te-
troxide in 0.2 M cacodylate solution for 1 h, and subse-
quently rinsed six times for 5 min in 0.2 M cacodylate
solution. After dehydration using a sequence of ethanol/
distilled water treatments for 10 min each, the specimens
were critical-point dried under CO2, coated with gold (300
A), and micrographs obtained with a Hitachi Model
S450A SEM.
Results
After 18 days and ten additions of fungal culture, H.
americanus embryos appeared healthy and free of fungal
infection. The visible organ anatomies and heartbeat rates
of the treated embryos were identical to those of the con-
trols. Scanning electron micrographs showed that the em-
bryonal surface was covered by an almost monoculture
of a rod-shaped bacterial strain (Fig. 1A). Some of the
embryos, at different stages of development, were occa-
sionally found to have very sparse coverage by three mor-
phologically different bacteria (Fig. IB), in addition to the
rod-shaped strain. Older embryos, near hatching, were
consistently observed to possess dense coverage by an al-
most monoculture of the rod-shaped bacterium (Fig. 1C).
Replicate inoculations of embryo homogenates on
Marine Agar plates resulted in the isolation of a maximum
of four, but usually fewer, strains of bacteria. One of the
strains (SGT-76), which was consistently isolated, was in-
hibitory to L. callinectes. This bacterium, a Gram-negative
rod insensitive to penicillin, was a pale salmon-colored
strain, and it was extremely slow growing on agar plates
and in liquid medium. For reasons unknown, the pH of
the culture medium seemed to rise to 8 or more during
fermentation. This rise in pH could provide a possible
explanation for the poor growth observed. Scanning elec-
tron micrographs of this bacterium, grown on millipore
filters, showed that the cells were identical in size and
shape, and had an identical growth pattern to those ob-
served on the surface of the natural embryos (Fig. ID).
An antifungal compound produced by the bacterial
strain grown in liquid medium was isolated and identified
as tyrosol, 4-hydroxyphenethyl alcohol (Fig. 2). The active
compound was isolated as a viscous oil which showed the
following spectral characteristics: IR (film): 3400, 3150
cm"1; HRMS requires 138.04 forC8H,0O2, found 136.06;
'H NMR, 200 MHz (acetone-do): 7.0 (d, 2 H, J = 8.6
Hz), 6.6 (d, 2 H, J = 8.6 Hz), 3.7 (t, 2 H), 2.7 (t, 2 H);
H. AMERICAM'S EMBRYOS PROTECTED BY BACTERIA
107
\
Figure 1. Scanning electron micrographs of healthy embryos of Homarus americanus under various
conditions. (A) Surface of embryo after exposure to the fungus Lagenidium callinectes. illustrating the lack
of fungal attachment ( 1250X). (B) Surface of embryo showing the coverage by the colonial rod morphotype
and the other three types occasionally found ( 1000.' ). (C) Surface of embryo, at near full gestation, showing
extensive and thick coverage by the rod-type bacterium (5000X). (D) Micrograph of the pure bacterium
grown on a millipore filter (2000X).
I3C NMR, 200 MHz (acetone-d6): 155.0, 131.0, 130.0,
1 1 5.5, 1 1 5.4, 63.8, 39. 1 (Fig. 2). All chemical and spectral
data were identical to those from the commercially-avail-
able 4-hydroxyphenethyl alcohol (tyrosol, Aldrich
#18,825-5). Tyrosol effectively inhibited growth of L. cal-
linectes in liquid culture at a concentration of 10 meg/
ml. In agar plate assays, 100 meg tyrosol per disk resulted
in an 8 mm zone of fungal inhibition.
Discussion
All the embryos observed were covered largely by a
single, rod-shaped bacterium, distinguishable from other
types by its characteristic dense, mosaic-like growth pat-
tern. The oldest embryos had the thickest coverage by this
particular strain, thus health and successful development
appear to be related to the degree of bacterial coverage.
Harper and Talbot (1984), who investigated embryos of
several Homarus species to determine if the presence of
epibiotic bacterial flora was related to loss of embryos
from the pleopods, also observed four bacterial morpho-
types. Their bacteria appear to be morphologically iden-
tical to those described in this article, including one that
they described as a "colonial rod." They found that em-
bryos from wild born and wild spawned H. americanus
were heavily covered by bacterial rods, and that these em-
bryos were successfully retained by the adult until hatch-
ing. Based upon the repetitive isolation of the salmon-
colored bacterium (SGT-76) from healthy embryos, and
its highly characteristic mosaic growth pattern on natural
108
M. S. GIL-TURNES AND W. FENICAL
HO
OH
4-hydroxyphenethyl alcohol
"tjrosol"
Figure 2. Chemical structure of the antifungal metabolite 4-hydrox-
yphenethyl alcohol.
surfaces and on filters, we believe that this bacterium is
the natural epibiont of Homams embryos. At the same
time, we recognize that this proposal will be difficult to
rigorously prove.
The antifungal compound produced by bacterium
SGT-76, 4-hydroxyphenethyl alcohol, or tyrosol. has pre-
viously been reported as a natural product from two fungal
species that are involved in symbiotic associations with
plants (Stoessl, 1969; Claydon et til., 1985). In an appar-
ently similar adaptation, those fungi seem to protect their
hosts against invasion by pathogenic fungi.
Protection of embryos by epibiotic bacteria has been
shown previously (Fisher, 1983; Gil-Turnes et a!., 1989)
for the caridean shrimp Palaemon macmdactylus, also an
external brooder. In the present study, the association of
H. arnericcmus embryos with a Gram-negative, rod-shaped
bacterium suggests a similar adaptation in which a vul-
nerable host is protected against pathogenic microorgan-
isms by symbiotic bacteria. At least one explanation for
the resistance of Homams embryos is the bacterial pro-
duction of the antifungal compound tyrosol in nature.
Although tyrosol is only a moderately potent antifungal
agent, the dense bacterial coverage observed would easily
result in high levels of the compound at the embryo sur-
face. Thus, tyrosol could function effectively to reduce
fungal encroachment.
The antifungal agents isolated from crustacean-asso-
ciated bacteria to date (tyrosol and 2,3-indolinedione from
a Palaemon macrodactylus-associaled bacterium), are
simple molecules with only modest potencies. These mol-
ecules appear to be unusually effective against Lageni-
dium, however, perhaps suggesting that they are targeted
to this and related fungal pathogens.
Studies of the bacterial symbionts of commercially im-
portant marine invertebrates could provide important in-
formation leading to the control of disease under aqua-
culture conditions. Indeed, there is a significant need to
develop inexpensive and environmentally safe antifungal
agents for this specific application. The simple molecules
discussed here, which appear to be derived from the com-
mon amino acids tyrosine and tryptophane, should be
considered in this application.
Acknowledgments
We would like to thank Ashley Yudin, William S.
Fisher, Wallis H. Clark, and Prudence Talbot for their
assistance in this study. The staff and students at the Bo-
dega Marine Laboratory were extremely helpful and their
cooperation is gratefully acknowledged. The work is
mainly a result of research sponsored by NOAA, National
Sea Grant College Program, Department of Commerce,
under grant #NA85AA-D-SG140, project number R/MP-
39, through the California Sea Grant Program, and in
part by the California State Resources Agency. Portions
of this research were also supported by the National In-
stitutes of Health, NCI, under grant CA44848. The U. S.
Government is authorized to reproduce and distribute for
governmental purposes.
Literature Cited
Claydon, N., J. F. Grove, and M. Pople. 1985. Elm bark beetle boring
and feeding deterrents from Phomopsis oblonga. Phytochemistry
24(5):937-943.
Cobb, J. S., and D. Wang. 1985. Fisheries biology of lobsters and cray-
fishes. Pp. 168-248 in The Biology of Crustacea. Vol. 10, Academic
Press. New York.
Fisher, \V. S. 1983. Eggs of Palaemon macmdactylus: II. Association
with aquatic bacteria. Biol. Bull. 164: 201-213.
Fisher, W. S., E. H. Nilson, L. F. N. Follett, and R. A. Shleser. 1976.
Hatching and rearing lobster larvae (Homams americamis) in a disease
situation. Aquaculture 1: 75-80.
Gil-Turnes, M. S., M. E. Hay, and \V. Fenical. 1989. Symbiotic marine
bacteria chemically defend crustacean embryos from a pathogenic
fungus. Science 246: 117-118.
Harper, R. E., and P. Talbot. 1984. Analysis of the epibiotic bacteria
of lobster (Homams) eggs and their influence on the loss of eggs from
the pleopods. Aquaculture 36: 9-26.
Nilson, E. H., W. S. Fisher, and R. A. Shleser. 1976. A new mycosis
of larval lobster (Homams americamis). J. Inverlebr. Pathol. 27: 177-
183.
Provenzano, A. J., Jr. 1985. Culture of crustaceans: general principals.
Pp. 279-280 in The Biology of Crustacea, Vol. 10. Economic Aspects:
Fisheries Culture. Academic Press, New York.
Stoessl, A. 1969. 8-Hydroxy-6-methoxy-3-methylisocoumarin and
other metabolites of Ceratocyslis fimbriata. Biochem. Biophys. Res.
Comm. 35(2): 186-191.
Reference: Biol. Bull 182: 109-1 16. (February, 1992)
Are Temperature and Photoperiod Necessary Cues for
Encystment in the Marine Benthic Harpacticoid
Copepod Heteropsyllus nunni Coull?1
JUDY WILLIAMS-HOWZE AND BRUCE C. COULL
Department of Biological Sciences and Belle H'. Banich Institute for Marine Biology & Coastal
Research, University of South Carolina, Columbia, South Carolina 29208
Abstract. Heteropsyllus nunni is a marine copepod that
builds a cyst and dwells within it during a period of ex-
tended diapause. The field abundance of this copepod has
been monitored for 10 years, but nothing is known about
the cues that induce and terminate encystment. In the
laboratory, different photoperiods and temperatures were
tested for their effects on encystment and excystment.
The photoperiod and temperature cues tested neither
induced nor inhibited encystment in H. nunni. Encyst-
ment occurred in all treatments, regardless of temperature
or photoperiod, suggesting that internal genetic cues, tied
to a specific ontogenetic stage, must be the central causal
factor. Copepods in the hot treatments encysted and ex-
cysted more rapidly than in the cold. Many copepods in
the cold treatment encysted (though later than copepods
in the hotter treatments), and most were still within the
cyst at the end of the 23-week experiment. There were
significantly more males within the full cysts than females.
A concurrent field study confirmed the known seasonal
patterns in the number of encystments relative to the
number of free-living forms; i.e., encystment took place
in the summer.
Introduction
A state of dormancy or diapause sometime during de-
velopment is a common adaptation for a myriad of
aquatic, terrestrial, and aerial invertebrates. Many of these
invertebrates have developed specialized adaptations that
protect against periodic (cyclic or acyclic) harsh environ-
mental conditions, such as dry seasons and extreme high
Received 22 July 1 99 1 : accepted 31 October 1991.
1 Contribution No. 892 from the Belle W. Baruch Institute for Marine
Biology & Coastal Research.
or low temperatures. Diapause and quiescence are two
such adaptations. Quiescence is characteristically brief,
irregular, and controlled by the effective adverse factors.
For example, a cold shock might send an invertebrate
into a state of quiescence; i.e., the animal enters and re-
mains in a state of torpor until the temperature rises,
causing normal physiological functions to resume. Qui-
escence is reversible, not fixed to a specific ontogenetic
instar, and may be induced repeatedly in the same indi-
vidual (Andrewartha, 1952; Tauber <??a/., 1986).
In contrast to quiescence, diapause interrupts the nor-
mal metabolic program away from its developmental
pathway at a specific ontogenetic stage. Moreover, dia-
pause is not controlled by the direct action of sporadic
environmental factors; rather it is cued in advance by
some predictable cyclic change in the environment (An-
drewartha, 1952; Danks, 1987). Diapause is, by definition,
neurohormonally driven (Danks, 1987) and involves a
more complicated developmental process that commits
the organism to a greater metabolic investment (Tauber
el ill., 1986). As an alternative to the normal develop-
mental pathway, diapause is favored when the expectation
of fitness accruing from active growth and reproduction
is less than that from survival in diapause (Cohen, 1970).
Insects are the most extensively studied of the dia-
pausing invertebrates (Tauber and Tauber, 1970; Tauber
et al, 1986; Danks, 1987). But copepods also exhibit di-
verse forms of diapause (Elgmork, 1980; Marcus, 1980;
Coull and Grant, 1981; Hairston, 1987), the physical
manifestation of the process varying with the order. Of
the major free-living orders, the largely planktonic Cal-
anoida produce primarily resting eggs (chitin-covered and
desiccant-resistant), although some species ofCalanus and
Neocalanus diapause in deep waters as a fifth stage (C-V)
109
110
J. W1LLIAMS-HOWZE AND B. C. COULL
copepodite (Miller el al., 1991). Diapausing eggs are
heavier than the subitaneous (immediately hatching) eggs
and will sink to the sediment. They are crucial in the
reproductive success of these copepods, as an aid in pred-
ator avoidance (Hairston, 1987) and as a mechanism for
surviving desiccation during drought (Taylor el al., 1990).
In the Cyclopoida, the individual copepod enters into
a state of dormancy at the fourth or fifth copepodite stage
(C-IV or V); rarely are resting eggs produced (reported
only for Mesocyclops edax, Wyngaard, 1988). The dia-
pausing cyclopoids sink to the bottom and remain ob-
scured by mud and detrital coverings (Fryer and Smyly,
1954; Elgmork, 1980; Nilssen, 1980).
Diapausing harpacticoids construct and reside in cysts,
typically at the adult stage, but diapause is not common
in the Harpacticoida. Most encysting species inhabit fresh
water and encyst during summer months (Sarvala, 1979;
Coull and Grant, 1981; Nalepa, 1985). Indeed, until the
discovery of encysted Heteropsyllus nunni Coull (Family
Cletotidae) in the marine environment, only freshwater
species of the family Canthocamptidae were thought to
encyst (Coull and Grant, 1981).
Timing for entering diapause is critical and is integrally
linked to reproductive success in copepods (Cohen, 1967,
1970; Taylor, 1980; De Stasio, 1990). This timing is most
often cued by photoperiod, temperature, or a combination
of both. The production of diapausing eggs by calanoids
appears to be controlled by the combined effects of pho-
toperiod (Marcus, 1980, 1982a, b) and temperature
(Cooley, 1978; Hairston and Olds, 1984; Marcus, 1987);
production may also differ with geography (Hairston and
Olds, 1984: Marcus, 1987). In cyclopoids, photoperiod is
the cue for induction of diapause (Watson and Smallman,
1971a, b; Elgmork and Nilssen, 1978). The dormant cy-
clopoid C-IV or C-V stages either overwinter or oversum-
mer, depending on the effect of geography on induction
(Elgmork, 1955; George, 1973; Cooley, 1978; Elgmork
and Langeland, 1980).
Little is known about the environmental cues that ini-
tiate diapause in harpacticoids. Sarvala (1979) determined
that a particular combination of photoperiod and tem-
perature were needed to induce encystment and excyst-
ment in freshwater Canthocampus staphylinus. Changes
in this light-temperature regime arrested development of
the copepodites, inhibited egg production in mature fe-
males, and induced the copepods to produce pre-diapause
oil droplets. Exposure to low temperature inhibited en-
cystment in C. staphylinus.
Heteropsyllus nunni is a recurring member of the in-
tertidal meiobenthos on South Carolina sandflats (Coull
and Grant, 1981). The number of free-living animals, rel-
ative to encysted ones, has been monitored for 10 years
(Coull, unpub.), but little is known about the biology of
this animal related to encystment. The objective of this
study was to experimentally investigate the effect of the
environmental cues, temperature, and photoperiod,
(known to induce diapause in other copepods), on timing,
sex ratio, ontogenetic stage specificity, and the number of
individuals encysting for the marine harpacticoid copepod
Heteropsyllus nunni.
Materials and Methods
Laboratory environmental cues experiment
Large numbers of Heteropsyllus nunni were collected
during January and February, 1990, from an intertidal
sand flat at Oyster Landing, North Inlet Estuary, South
Carolina, USA (33° 19.0' N, 79° 11.6' W). At random
sites along the exposed sand flat during low tide, the upper
two centimeters of sand containing the copepods were
scraped up by hand, placed in a bucket with seawater and
transported back to the laboratory in Columbia, South
Carolina. Live animals were extracted from the sand using
a 3% solution of isotonic magnesium chloride. The mag-
nesium chloride solution was added to small amounts of
sand, shaken well, and, within 10 min, all living meiofauna
in the sand were anesthetized and then decanted. H. nunni
were separated from other meiofauna under a stereo dis-
secting microscope and placed in large petri dishes con-
taining filtered artificial seawater (ASW). All copepods
were held in an incubator at 18°C, with 16:8 h day:night
cycle until sufficient nauplii had hatched. A sand substrate
for culture was prepared as follows. Clean sand (300-500
p.m size fractions; obtained from the sand flat at Oyster
Landing) in a 500 ml flask, was autoclaved, covered with
F/2 medium solution (Guillard, 1972), and autoclaved
again for 10 min. The sand and medium were then in-
oculated with 10 mis of Phaeodactylum tricornutum.
Within a week, Phaeodactylum was growing on the sand
grains, providing an adequate grazing substrate for //.
nunni. The culture dishes could then be prepared as fol-
lows. The substrate of sand and algae was removed from
the flask with a sterile pipet, washed with filtered seawater
to remove the excess culture medium and placed in sterile
plastic petri dishes with sterile-filtered ASW (salinity 29-
30%o). Nauplii were held in 35 X 10 mm size petri dishes,
and all other stages were held in 60 X 20 mm size dishes.
All experimental dishes with copepods were established
during the same day.
The experiments were conducted in two incubators,
one set at the ambient winter temperature ( 10°C) in North
Inlet, South Carolina, and one set at the ambient early
summer temperature (20°C). Early summer (May, June)
is the time of encystment for H. nunni in the field (Coull
and Grant, 1981). Within these incubators were black
boxes within which a long night:short day ( 15 h dark:9 h
light) regime could be simulated in isolation from the
other portion of the incubator which was set for long day:
CUES FOR ENCYSTMENT IN H. NVNNI
111
short night ( 1 5 h light:9 h dark). Thus, the four treatment
conditions were: cold-short day (Cold-SD), cold-long day
(Cold-LD). hot-short day (Hot-SD), and hot-long day
(Hot-LD).
The black boxes, each with a hinged lid, were con-
structed of thick (1.85 cm) styrofoam. A slit was cut in
the top, and a time-controlled fluorescent light was placed
above the slit. The light and box were completely covered
with a double layer of black cloth, and cardboard was
taped under the upper black box separating it from the
lower open light source. We tested the black box for pos-
sible light entrance by placing a 35 mm camera containing
400 ASA film in the box and exposing the film using a
delayed automatic timer for 6 exposures. The same cam-
era and film were then taken to a photographic darkroom
and another six frames were exposed using the same times
and settings. There were no differences between the frames
exposed within the black box and those in the darkroom,
indicating no significant light leakage into the boxes.
Replicate sets of life-history stages (Table I), chosen to
represent a wide range of age classes, were exposed to the
environmental conditions constituting each experimental
treatment. Each ontogenetic stage was isolated in a sep-
arate petri dish so that their growth and encystment could
be compared. The number of copepods in each dish was
determined by the availability of that life-history stage at
the beginning of the experiment.
Nauplii, copepodites, or adults obtained from the field
might have been pre-cued by their environments to encyst
before the initiation of the experiment. To preclude this,
gravid females and females with ovaries full of egg masses
were used in each treatment so that hatching nauplii would
be exposed only to the temperature and photoperiod re-
gime specified by the experimental protocol.
The photoperiod timers in the incubators were coor-
dinated so that daylight would occur in all treatments
from 12 noon to 6 pm daily, allowing all feeding, changing
of water, observations, and counts to be made during this
"daylight on" time. All copepods were fed concentrated
drops of the alga Isochrysis sp., or additional sand with
P. tricornutum (as needed). Water was changed at least
once a week, more frequently in the smallest dishes. Excess
algal clumps, feces, and detritus were removed by pipet
weekly. For each dish, weekly counts were made of mor-
tality, the number of full cysts and empty cysts, the stage
of development (of nauplii), the number of females with
eggs, females with developing eggs in their ovaries, and
mating pairs.
The experiment was initiated on March 15. 1990, and
was terminated on August 19, 1990; 23 weeks. At the end
of the experiment, all of the dishes were removed from
the incubators, and 10% formalin with Rose Bengal was
added to each dish to preserve all copepods. Every indi-
vidual was counted and categorized as to life-history stage
Table I
Number of individuals representing each of five life-history slages
of copepods fUeteropsyllus nunnij placed within each
of four experimental treatments"
No.
Replicate
Stage
in dish
dishes
Nauplii
20
5
Male
5
3
Females (no eggs; full ovaries)
5
3
Gravid females (with eggs attached)
5
3
Post-gravid females'1
4
3
a The experimental treatments are denned in the legends to Table II
and Figure 1.
b Females removed from isolated dishes that contained nauplii.
(nauplii, copepodite, adult, male, female). Each cyst was
noted as being full or empty, and, for all full cysts, the
copepod was removed, dissected, and sexed.
Field population study
A field study was conducted on the same intertidal sand
flat at Oyster Landing, North Inlet, South Carolina (USA)
from which H. nunni had been obtained for the laboratory
experiments.
Quantitative collections were made by random hand
coring with a 2.54 cm diameter core tube in the upper 10
cm of sediment during low tide. Eight samples were taken
monthly for one year (Sept. 1988-Aug. 1989). All samples
were immediately preserved with 10% buffered Formalin
with Rose Bengal added. In the laboratory, copepods were
extracted via elutriation where sand was placed in a sep-
aration flask and water was gently bubbled up through
the sand. This loosened and released the copepods from
the sand grains, allowing them to be captured in the out-
flowing water. Individuals of//, nunni were counted, sexed
and life-history stage recorded.
Statistical analysis
Free-living and encysted copepod abundance within
the four experimental treatments was analyzed separately
by the General Linear Model (GLM) procedure (1-way
ANOVA, treatment vs. final abundance), and Tukey's
multiple comparison procedure of SAS to compare treat-
ment effect (SAS Institute, 1985). Data for all of the life-
history stage within a treatment were pooled, because the
developmental rates of the representatives of each stage
were indistinguishable.
Field data on free-living copepods (males and females)
and encysted copepods were loglo (n + 1 ) transformed to
meet the assumptions of normality and homoscedasticity.
The seasonal abundance of H. nunni in the field, by
112
J. WILLIAMS-HOWZE AND B. C. COULL
en
a
O
a.
LU
a.
O
u
(C
111
CD
2
LU
5
HOT-LD HOT-SD COLD-LD COLD-SD
TREATMENT
Figure 1. The effect of temperature and photopenod regimes on the
mean number of free-living copepods (i.e.. not encysted). The duration
of the experiment was 23 weeks. Treatments: Cold, 10°C, Hot, 20°C;
LD (long day). \5 h light:9 h dark; SD (short day). 9 h light: 15 h dark.
Error bars are one standard deviation of mean.
O
u.
O
cc
LU
CO
<
LU
HOT-LD HOT-SD COLD-LD
TREATMENT
COLD-SD
Figure 2. The effect of temperature and photopenod regimes on the
mean number of full and empty cysts. The duration of the experiment
was 23 weeks, and treatment conditions are as listed in the legend to
Figure 1 . Error bars are one standard deviation of mean.
month, as well as the number of females compared to
males, were also compared using the GLM procedure ( 1-
way ANOVA, month vs. total) and Tukey's multiple
comparison procedure. All significance levels were set at
alpha < 0.05.
Results
Laboratory environmental cues experiment
A. Free-living Heteropsyllus nunni. The final mean
numbers of free-living H. nunni in the four experimental
treatments were not significantly different (P = 0.50; 1-
way ANOVA, final number of free-living animals vs.
treatment). Although the total mean in the Hot-SD treat-
ment was slightly more than double the mean in the Hot-
LD (Fig. 1 ), the great variability within treatment masked
any significant difference. There were no significant dif-
ferences between the number of free-living males com-
pared to free-living females among treatments, again due
to high variability within treatment.
B. Encysted H. nunni. Sixteen culture dishes within
each treatment represented five different life-history stages.
The frequency of encystment events (at least one cyst in
a dish) was surprisingly high (69%) in all four treatments.
Frequency of encystment events in all dishes (16 total)
by treatment was: Hot-SD = 11/16; Hot-LD = 11/16;
Cold-SD = 11/16; Cold-LD = 12/16, indicating copepods
encysted in most of the five life-history stages originally
placed within the dish. The copepods within the cysts
were all C-VI, unmated adults that had developed from
nauplii in each dish (regardless of the ontogenetic stage
placed in the culture dishes). There were no reproductive
or post-gravid females, no mated males, and no stages
younger than C-VI encysted.
There were significantly more empty cysts compared
to full cysts at the end of the experiment (P = 0.018) over
all four treatments (Fig. 2). Empty cysts in Hot-SD (x
= 6.3), Cold-LD (x = 0.75), and Cold-SD (x = .37) con-
ditions were significantly different from each other, but
Hot-SD and Hot-LD (x = 3.6) were not (Tukey's multiple
comparison test. P < 0.05). There were 48 encystment
events in the Cold-LD treatment, 51 in the Cold-SD
treatment, 81 in the Hot-LD treatment, and 117 in the
Hot-SD treatment, but there was no significant difference
(P = 0.39) in mean numbers of cysts between treatments
(Fig. 2).
The time to first encystment for the Hot treatments
was 37 days and 67-77 days for the Cold treatments (Table
II); thus, although encystment was delayed in the cold, it
was not inhibited. This delay resulted in more cysts with
the copepod still inside compared to the Hot treatments,
thus, more full cysts than empty cysts at the termination
of the experiment (Fig. 2). The number of full cysts among
treatments was not significantly different, due, again, to
high dish to dish variability.
C. Proportion of males to female H. nunni in cysts.
There were significantly more (P = <0.001) males than
Table II
Time to first encvstment for nauplii in four experimental treatments
Date 1st
Days from
No. cysts
Begin
encvstment
nauplii to
1 st date
Treatment*
date
observed
encystment
observed
Cold-long dav
3/15/90
6/1/90
77
3
Cold-short dav
3/15/90
5/23/90
67
2
Hot-long day
3/15/90
4/23/90
37
35
Hot-short day
3/15/90
4/23/90
37
9
* Treatment: Cold, 10°C; Hot, 20°C; long day. 15 h light:9 h dark;
short day, 9 h light: 1 5 h dark.
Numbers of cysts are totals found on 1st date of encystment. All rep-
licates are combined for each treatment.
CUES FOR ENCYSTMENT IN H. NUNNl
113
females in cysts in all treatments (Fig. 3); mean male/
female ratio = 3.5/1.
Field population study
A. Number of free-living H. nunni/numher of cysts over
12 months. There was a significant difference (P = <0.00 1 )
between mean copepod abundance by month (Fig. 4).
January and February had significantly more free-living
H. nunni than other months (April-November); March
was not significantly different from Jan-Feb or Apr-Nov
(Tukey's multiple comparison procedure). Free-living H.
nunni reached maximum abundance in winter and were
low in number, then absent as summer progressed. The
mean number of full cysts throughout the year was not
significantly different between months, because the num-
ber of cysts in the cores was extremely low (Fig. 4). Cysts
were most abundant in summer, when free-living //. nunni
were absent from the core samples (Fig. 4).
B. Free-living males and females over 12-month study.
The mean number of males compared to females was
significantly different over the one-year sampling period
(males and females both with P = <0.001) (Fig. 5). The
number of males was slightly greater than females in Oc-
tober and November (time of emergence from cysts). The
population was dominated by females from December to
April, the period of peak egg production (Fig. 5). Free-
living males and females disappeared in summer during
peak encystment time (May, June, July).
Discussion
In the field, H. nunni encysts in early summer (day-
length 14 h, temperature 15-18°C). In the laboratory,
therefore, we expected H. nunni not to encyst under winter
(i.e.. cold-short day) conditions. Nevertheless, encystment
occurred in all treatment conditions and was not inhibited
cc
HI
m
5
<
LLJ
HOT-SD COLD-LD
TREATMENT
COLD-SD
Figure 3. The effect of temperature and photopenod regimes on the
mean number of male compared to female copepods after removal from
the full cysts. The duration of the experiment was 23 weeks, and treatment
conditions are as listed in the legend to Figure I . Error bars are one
standard deviation of mean.
SEPT OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG
MONTHS
Figure 4. Mean number of encysted and free-living copepods taken
from the cores during the field study. Core samples were taken once a
month for twelve months. Number per 10 cm- is the unit of density for
meiobenthos, in contrast to number per m: used for macrobenthos. Error
bars are one standard deviation of mean.
by the dark, cold environment (9 h light at 10°C). En-
cystment in H. nunni must be genetically induced, because
sexually immature adults encysted regardless of the sur-
rounding temperature or photoperiod regime. Our results
are in direct contrast to those of previous research, which
indicate that photoperiod and temperature are necessary
mechanisms for inducing copepod diapause, e.g.. for cal-
anoids (Marcus. 1980, 1982a, b, 1987; Hairston et ai.
1990), cyclopoids (Watson and Smallman, 197 la, b; Elg-
mork and Nilssen, 1978). and freshwater harpacticoids
(Sarvala, 1979). Additionally, no female H. nunni dis-
sected from cysts had attached spermatophores, egg sacs,
or maturing ova, nor did any males have developing sper-
matophores. Because mated adults would show at least
some of these characteristics, the encysted individuals
must not have mated. In Canthocamptus staphylinm,
however, females with attached spermatophores encyst
(Sarvala, 1979), and fertilized, adult females of Cyclops
strenuus diapause (Naess and Nilssen, 1991).
Although temperature and photoperiod apparently did
not specifically cue encystment, they did affect the de-
velopmental rates of//, nunni. The most significant effect
was on nauplii, because naupliar development to adult,
and then to encystment, took twice as long in cold treat-
ment (67-77 days) as it did in the hot treatments (37
days) (Table II). In the field, H. nunni mate and produce
eggs during the winter months. Nauplii hatch from the
eggs in late winter or early spring (March, April) when
temperatures in the estuary are still quite cool. Therefore,
the cold treatments were probably closer to the normal
field conditions in temperature and early naupliar devel-
114
J. WILLIAMS-HOWZE AND B. C. COULL
O
o
6
30
25
20
15
10
SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG
MONTHS
Figure 5. Mean number of free-living male and female copepods taken from the cores during the field
study. Core samples were taken once a month for twelve months. Number per 10 cnr is the unit of density
for meiobenthos, in contrast to number per rrr used for macrobenthos. Error bars are one standard deviation
of mean.
opment than the hot treatment regime. Copepodites nor-
mally reach adulthood in early summer (April-May) in
the field, and encyst during summer months only (Fig.
4). Greater total number of cysts in the hot treatments
(81 Hot-LD, 117 Hot-SD) versus the cold treatments (48
Cold-LD, 5 1 Cold-SD) were probably due to increased
rates of development. The high number of encystment
events in the hot-short day was unexpected, particularly
because longer photoperiod has been implicated as the
main cue triggering summer dormancy in other copepods
(Watson and Smallman, 197 la, b; Sarvala, 1979).
Not all of the sexually immature adult H. nunni en-
cysted. In most treatment dishes there were mating and
reproducing free-living copepods throughout the entire
23 weeks, along with encysted individuals; this was un-
expected, because no free-living forms have been found
in the summer (Coull and Grant, 1981, and Fig. 5). Coull
and Grant (1981) hypothesized that the free-living pop-
ulation either moved to another area, or all members en-
cysted. The calanoid copepod Diaptomus sangiiineus
produces diapausing and subitaneous eggs sequentially
during the same reproductive period, and Hairston and
Munns (1984) suggested that it was using a bet-hedging
strategy (sensu Stearns, 1976), anticipating that an envi-
ronmental catastrophe would not occur or would be less
severe than expected. Reproductive success could then be
insured in either situation. In harpacticoid and cyclopoid
copepods, such a bet-hedging strategy is generally not used,
because the diapausing stage is not an egg, but an indi-
vidual (i.e., either copepodite or adult). If the adult is the
diapausing organism insuring reproductive success (as
opposed to dispersed diapausing eggs), a bet-hedging
strategy would not be expected (Hairston, 1987). However,
we found free-living harpacticoids along with the encysted
ones, as did Cole (1953) and Sarvala (1979). Perhaps these
free-living forms are also bet-hedgers, taking the chance
that they will not be negatively affected in their non-dia-
pause state. Our inability to find such proposed bet-hedg-
ers in the field (i.e., free-living H. nunni in the summer)
may be a function of them occurring in very low abun-
dance.
In our laboratory experiment there were consistently
more males than females, both in cysts and free-living.
These findings are very different from those of Sarvala
(1979), who observed that Canthocamptus staphylinus
males were absent from cysts. However, in the cyclopoid
Cyclops victims and Thermocyclops crassits, more males
than females emerge from diapause (George, 1973, and
Maier, 1989, respectively). The initial data on H. nunni
(Coull and Grant, 1981) indicated a female to male ratio
in the cysts of 2.3:1, but over an 1 1 -year sampling period,
the female-to-male ratio for free-living H. nunni was 1.6:
1 (Coull and Dudley, 1985). Males within cysts outnum-
bered females by at least 2:1 in the laboratory (Fig. 3).
For copepods, a sex ratio other than 1 : 1 indicates a shift
in sexual selection pressure. Male dominance in this ex-
periment may be a laboratory effect, as excessive homo-
zygosity leads to shifting of the sex ratio in favor of males
(Hicks and Coull, 1983). An imbalanced ratio could be
due to homogeneity of the environment (i.e., small culture
dishes), which favors inbreeding, and results in a more
homogeneous population. Population density can also
CUES FOR ENCYSTMENT IN H NUNNI
115
influence sex ratios. Hicks (1984) found that male Par-
astenhelia megarostnim dominated only when the pop-
ulation density was high; in lower densities, females dom-
inated. Another potential influence of gender density is
"sexual switching." Hicks and Coull (1983) cite reports
of genetic males becoming phenotypic females in response
to low population density. In our study, H. nunni males
dominated over females during Sept-Nov (low density
population, 4 per 10 cm2). In December, the female pop-
ulation increased rapidly from 2 to 23 per 10 cm2, but
the male population remained at previous abundances
(Fig. 5). If there were no sex-switching, perhaps this phe-
nomenon was related to developmental differences be-
tween males and females.
In certain harpacticoids, males mature much faster than
females (Fleeger and Shirley, 1 990). Samples taken in early
spring had mostly males and copepodites (stages 4-5) and
the number of males within the population remained
constant; as the copepodites developed, more females ap-
peared, and eventually there were more females than
males. Perhaps a similar developmental sequence occurs
in H. nunni, where males develop, encyst, and excyst ear-
lier than the females, biasing the ratio towards males as
the copepods emerge from the cysts. As other individuals
mature (females), the ratio then switches to female dom-
inance (Fig. 5).
Biotic factors that induce diapause were not directly
tested, but two possibilities exist. Because H. nunni cysts
are not resistant to desiccation (they collapse around the
copepod and dry up when removed from water), perhaps
the cyst is used to avoid competition or predation. Where
H. nunni occurs in South Carolina, the five most abundant
copepods (80% of all copepods) have high maximum
densities (1056 per 10 cnr/per) and reproduce from sum-
mer through fall (Coull and Dudley, 1985). Heteropsyllus
nunni reproduces and reaches its maximum population
density in winter. In summer months, when other har-
pacticoids are at their peak, H. nunni is within its cyst,
dormant. Competition avoidance could possibly be in-
ducing the encystment diapause in H. nunni.
Large numbers of juvenile fish that selectively prey on
harpacticoid copepods (Ellis and Coull, 1989; Nelson and
Coull, 1989) occupy South Carolina estuaries in the spring
and summer. A female H. nunni carrying eggs is highly
visible. The egg sac is large (40+ eggs) and has a bluish
tint, and thus H. nunni is a susceptible prey item. By
reproducing in the winter when there are few juvenile
fish, H. nunni is less available to predation. As the abun-
dance of juvenile fish increases in the summer, H. nunni
encysts. H. nunni cysts are cryptic (Coull and Grant, 1981)
i.e., they are indistinguishable from the surrounding sand.
Such camouflage would seem efficient in avoiding visual
predators. While there is no field evidence that H. nunni
is sought as a prey item by young fish, only one fish
(Leiostomus xanthimis) that consumes mud dwelling
harpacticoids has been thoroughly studied from the locale
(Feller et al., 1990). Predation avoidance also could be
influencing the encystment diapause of H. nunni.
We have tested whether temperature and photoperiod
(generally important cues for copepod diapause) were sig-
nificant factors inducing encystment in H. nunni. Al-
though cold temperatures slowed development (increasing
time to encystment) and hotter temperatures accelerated
naupliar development (decreasing time to encystment),
photoperiod appeared to have no impact on development
or encystment. In the past, perhaps, temperature and
photoperiod were important environmental factors cuing
these copepods of an impending catastrophe. Now, how-
ever, the interactions that induce diapause may be so
evolved that the specific catastrophe that favored encyst-
ment in the past is obscure. We conclude that for H. nunni.
the encysted diapause state is a relic adaptive response
that has become internalized into a developmental ne-
cessity.
Acknowledgments
We gratefully acknowledge Dr. G. T. Chandler for as-
sistance in culturing H. nunni, Drs. R. J. Feller, S. E.
Stancyk, M. B. Thomas, and N. Watabe, Mrs. B. W.
Dudley, and two anonymous reviewers for constructive
comments on earlier drafts of this manuscript. This re-
search is part of a dissertation submitted by Judy Williams-
Howze as partial fulfillment of the requirements for the
Ph.D. degree in Biological Sciences at the University of
South Carolina. It was supported by grants in aid from
Sigma Xi, the Slocum-Lunz Foundation (JW-H) and the
Biological Oceanography section of the National Science
Foundation, Grant OCE 89-16255 (BCC).
Literature Cited
Andrewartha, H. G. 1952. Diapause in relation to the ecology of insects.
Biol. Rev. 27: 50-107.
Cohen, D. 1967. Optimizing reproduction in a randomly varying en-
vironment when a correlation may exist between the condition at
the time a choice has to be made and the subsequent outcome. J.
Theor. Biol 16: 1-14.
Cohen, D. 1970. A theorectical model for the optimal timing of dia-
pause. Am. Nat. 104: 389-400.
Cole, G. A. 1953. Notes on copepod encystment. Ecology 34: 208-
211.
Cooley, J. M. 1978. The effect of temperature on the development
diapausing and subitaneous eggs in several freshwater copepods.
Crustaceana 35: 27-34.
Coull, B. C., and J. Grant. 1981. Encystment discovered in a marine
copepod. Science 212: 342-344.
Coull, B. C., and B. Dudley. 1985. Dynamics of meiobenthic copepod
populations: a long-term study (1973-1983). Mar. Ecol. Prog. Ser.
24: 219-229.
Danks, H. V. 1987. Insect Dormancy: An Ecological Perspective. Biol.
Sur. Can. Mono. 1: 439 pp.
116
J. WILLIAMS-HOWZE AND B. C. COULL
DeStasio, B. T. 1990. The role of dormancy and emergence patterns
in the dynamics of a freshwater zooplankton community. Limnol.
Oceanog. 35: 1079-1090.
Elgmork, K. 1955. A resting stage without encystment in the annual
life cycle of the freshwater copepod Cyclops strenmts strenuus. Ecol-
ogy: 36: 739-743.
Elgmork, K., and J. P. Nilssen. 1978. Equivalence of copepod and
insect diapause. I'erli. Inter. Verein. Limnol. 20: 2511-2517.
Elgmork, K. 1980. Evolutionary aspects of diapause in freshwater co-
pepods. Pp. 4 1 1-4 1 7 in Evolution and Ecology of Zooplankton Com-
munities, C. Kerfoot, ed. University Press. New England.
Elgmork, K., and A. Langeland. 1980. Cyclops scutijer Sars — one and
two-year cycles with diapause in the meromictic lake Blankvatn. Arch.
HydrobioL 88: 178-201.
Ellis, M. E., and B. C. Coull. 1989. Fish predation on meiobenthos:
field experiments with juvenile spot (Leiostomus xanlhurus). J. Exp.
Mar. Biol. Ecol. 130: 19-32.
Feller, R. J., B. C. Coull, and B. Hentschel. 1990. Meiobenthic co-
pepods: tracers of where juvenile Leiostomus xanthurus (Pisces) feed?
Can J. Fish. Aquat. Sci. 47: 1913-1919.
Fleeger, J. W., and T. Shirley. 1990. Meiofaunal responses to sedi-
mentation from an Alaskan spring bloom. II. Harpacticoid population
dynamics. Mar. Ecol. Prog. Ser. 59: 239-247.
Fryer, G., and W. J. P. Smyly. 1954. Some remarks on the resting
stages of some freshwater cyclopoid and harpacticoid copepods. Ann.
Mag. Nat. Hist. 7: 65-72.
George, D. G. 1973. Diapause in Cyclops vicimts. Oikos 24: 136-142.
Guillard, R. L. 1972. Culture of phytoplankton for feeding marine
invertebrates. Pp. 20-60 in Culture ol Marine Invertebrate Animals,
W. L. Smith and M. H. Chanley. eds. Plenum Press. New York.
Hairston, N. G., Jr. 1987. Diapause as a predator-avoidance adaptation.
Pp. 282-290 in Predalion: Direct and Indirect Impacts on Aquatic
Communities, C. W. Kerfoot and A. Sih. eds. University Press. New
England.
Hairston, N. G., Jr., and W. R. Munns Jr. 1984. The timing ot copepod
diapause as an evolutionary stable strategy. Am. Nat. 123: 733-75 1 .
Hairston, N. G., Jr., and E. J. Olds. 1984. Population differences in
ihi. Inning of diapause: adaptation in a spatially heterogeneous en-
vironment. Oecologia 61: 42-48.
Hairston, N. G., Jr., T. A. Dillon, and B. D. DeStasio. 1990. A field
test for the cues of diapause in a freshwater copepod. Ecology 71:
2218-2223.
Hicks, G. R. F., and B. C. Coull. 1983. The ecology of marine meiob-
enthic harpacticoid copepods. Oceanogr. Mar Biol Ann. Rev. 21:
67-175.
Hicks, G. R. F. 1984. Spatio-temporal dynamics of a meiobenthic co-
pepod and the impact of predation-disturbance. J. Exp. Mar. Biol.
Ecol. 81: 41 '-72.
Maier, G. 1989. The seasonal cycle of Thermocyclops crassus (Fischer
1853) (Copepoda:Cyclopoida) in a shallow, eutrophic lake. Hydro-
biclogia 178: 43-58.
Marcus, N. H. 1980. Photoperiodic control of diapause in the marine
calanoid copepod Labidocera aestiva. Biol Bull 159: 31 1-318.
Marcus, N. H. 1982a. The reversibility of subitaneous and diapause
egg production by individual females of Labidocera aestiva (Cope-
poda: Calanoida). Biol. Bull 162: 39-44.
Marcus, N. H. 1982b. Photoperiod and temperature regulation of dia-
pause of Labidocera aestiva (Copepoda: Calanoida). Biol. Bull. 162:
45-52.
Marcus, N. H. 1987. Differences in the duration of egg diapause of
Labidocera aestiva (Copepoda: Calanoida) from the Woods Hole.
Massachusetts, region. Biol. Bull. 173: 169-177.
Miller, C. B., T. J. Cowles, P. H. Wiebe, N. J. Copley, and H. Grigg.
1 99 1 . Phenology in Calanus finmarchicus: hypothesis about control
mechanisms. Mar. Prog. Ser. 72: 79-91
Naess, T., and J. P. Nilssen. 1991. Diapausing fertilized adults. Oec-
ologia 86: 368-37 1 .
Nalepa, T. F. 1985. Occurrence of a resting stage in cyclopoid and
harpacticoid copepods in nearshore Lake Michigan. J. Great Lakes
Res 11: 59-66.
Nelson, A. L., and B. C. Coull. 1989. Selection of meiobenthic prey
by juvenile spot (Pisces): an experimental study. Mar. Ecol. Prog.
Ser 53:51-57.
Nilssen, J. P. 1980. When and how to reproduce: a dilemma for lim-
netic cyclopoid copepods. Pp. 4 1 8-424 in Evolution and Ecology of
Zooplankton Communities. C. W. Kerfoot, ed. University Press, New
England.
Sarvala, J. 1979. A parthenogenetic life cycle in a population of Can-
thocamplus staphylinus (Copepoda. Harpacticoida). Hydrobiologia
62: 113-129.
SAS Institute. 1985. SAS User's Guide: Statistics, Version 5th edition.
SAS Institute Inc., Cary, North Carolina.
Stearns, S. C. 1976. The evolution of life history traits: a critique of
the theory and a review of the data. Ann. Rev. Ecol. Syst. 8: 145-
171.
Tauber, M. J., and C. A. Tauber. 1970. Photoperiodic induction and
termination of diapause in an insect: response to changing day lengths.
Science 167: 170.
Tauber, M. J., C. A. Tauber, and S. Mazaki. 1986. Seasonal Adap-
tations of Insects. Oxford Univ. Press. Pp. 4 1 1 .
Taylor, F. 1980. Optimal switching to diapause in relation to the onset
of winter. Theor. Pop. Biol 18: 125-133.
Taylor, B. E., G. A. Wyngaard, and D. L. Mahoney. 1990. Hatching
of Diaptomus stagnalis from a temporary pond after a prolonged
dry period. Arch. HydrobioL 117: 271-278.
Watson, N., and B. N. Smallman. 1971a. The role of photoperiod and
temperature in the induction and termination of an arrested devel-
opment in two species of freshwater cyclopoid copepods. Can. J.
Zoo/. 49: 855-862.
Watson, N., and B. N. Smallman. 1971b. The physiology of diapause
in Diacyclops navus Herrick (Crustacea, Copepoda). Can. J Zoo/.
49: i449_i454.
Wyngaard, G. 1988. Variation in dormancy in a copepod: evidence
from population crosses. Hydrobiologia: 167/168: 367-374.
Reference: Biol. Bui/. 182: 1 17-128. (February, 1992)
The Nature and Origin of the Epidermal Scales of
Notodactylus handschini—VM Unusual Temnocephalid
Turbellarian Ectosymbiotic on Crayfish
from Northern Queensland
JOSEPH B. JENNINGS, LESTER R. G. CANNON1, AND ADRIAN J. HICK
Department of Pure and Applied Biology. Baines Wing. University of Leeds. Leeds LS2 9JT
England and ^Queensland Museum. PO Box 300. South Brisbane. Queensland 4101. Australia
Abstract. The temnocephalid Notodactylus handschini.
ectosymbiotic on the crayfish Cherax quadricarinatus
from northern Queensland, is unique among known tur-
bellarians in having its dorsal epidermis covered by rows
of closely adjacent scales. These are borne on epidermal
plinths separated by arthrodial gutters and are up to 100
^m tall with rhombic bases 40-55 Mm by 15-20 ^m.
Above the bases, the rhombic cross section gradually be-
comes oval so that the scales are essentially elongate con-
oids, the slender tips of which curve inwards towards the
worm's mid-line. In mature worms, the more median
scales may be reduced distally into squat truncated cones
only 40-50 ^m tall. The scales consist of glycoprotein;
rhabdites discharged from cells in the dorsal parenchyma
contribute the protein, whereas the carbohydrate com-
ponent probably comes from the glycocalyxes of the epi-
dermal microvilli. The latter act as templates around
which the glycoprotein mixture coalesces, seemingly by
a simple tanning process, into tightly packed tubes 180-
200 nm in diameter with walls 40-45 nm thick. The scales
lack any limiting wall or membrane other than a loose
amorphous layer, 90-150 nm thick, formed by disinte-
gration of the tubes distally and compensated for by
continuous growth basally. Each scale is attached to its
epidermal plinth by the bases of its constituent tubes en-
sheathing the microvilli; attachment is reinforced by cross-
striated fibrils, probably collagen, embedded in the epi-
dermis and inserted between the microvilli into tube bases
near the scales' corners. Scale surfaces bear rich growths
of microorganisms. The use of rhabdites to form per-
Received 7 August 1991; accepted 30 September 1991.
manent scales is probably an adaptation to the worm's
unusual sedentary habit; it supports, paradoxically, an
earlier hypothesis that the primary function of rhabdites
in turbellarians other than temnocephalids is to provide
a continuously renewable coating compatible with epi-
dermal ciliation.
Introduction
The epidermis in turbellarian flatworms (comprehen-
sively reviewed by Tyler. 1984) is typically a monolayered
ciliated epithelium, with microvilli, made up of distinct
cuboidal, squamous, or columnar cells. It can, though, be
syncytial or insunk with its nuclei and some cytoplasm
lying among or even below the subepidermal musculature.
The epidermis is penetrated by the necks of subepidermal
glands and the dendrites of sensory structures that pass
between or through the epithelial cells when these are
present.
These basic patterns are remarkably constant through-
out the Turbellaria and persist, for example, in those en-
toparasitic species that lack normal entodermal alimentary
systems and use the epidermis as their sole means of nu-
trient uptake. In the Fecampiidae, living in the hemocoel
of amphipod and isopod crustaceans, the epidermis re-
mains typically turbellarian. and the only apparent struc-
tural modification is an increase in the density and length
of the microvilli (Jagersten, 1942; Christensen and Kan-
neworff, 1964; Bresciani and K0ie, 1970; Blair and Wil-
liams, 1987). The fecampiids take up only soluble nu-
trients, but even in species where the epidermis actively
secretes digestive enzymes and takes up paniculate ma-
terial for completion of digestion intracellularly. as in the
117
118
J. B. JENNINGS ET AL.
rhabdocoel Acholades asteris living in the tube-feet of
starfishes, the cells remain columnar, ciliated, and tra-
versed by the necks of subepidermal glands (Jennings,
1989).
The only major departure from the typical turbellarian
pattern occurs in the Temnocephalida, which are ecto-
symbiotes of freshwater decapod crustaceans and a few
other hosts. In most temnocephalids, cilia are restricted
to small areas of the tentacles or around the excretory
pores, locomotion is by muscular looping using the ten-
tacles and simple posterior sucker and not by ciliary glid-
ing, and the syncytial epidermis is bounded distally by a
narrow, clear zone of vesicular epitheliosomes (Williams,
1975. 1980, 1986). The epidermal surface still bears mi-
crovilli, though, and the syncytium is honeycombed by
numerous cell necks through which subepidermal glands
discharge their secretions.
The most extreme epidermal modification in the Tem-
nocephalida and, indeed, in the Turbellaria as a whole so
far as is presently known, occurs in Notodactylus hand-
sc/iini, an ectosymbiote of various crayfishes from Papua
New Guinea and northern Australia. In this species, the
entire dorsal surface is covered by golden-brown scales
that are much taller than the underlying epidermis (Baer,
1945, 1953; Cannon, 1991 ). These have not yet been de-
scribed in any detail, and we report here, therefore, on
their nature, origin, and mode of formation, as part of a
wider study on the general biology of this unusual tur-
bellarian.
Materials and Methods
Adults, juveniles, and hatchlings of Notodactylus
handschini (Baer 1945) (Turbellaria: Temnocephalida)
were collected from the lateral margins of the carapace
of the freshwater decapod crustacean Cherax quadricar-
inatus (von Martens 1868), a northern Queensland species
held in culture in farm ponds near Gympie, southeast
Queensland. Specimens for histological and histochemical
studies were fixed in Bouin's fluid, 90% ethanol or 10%
formalin buffered to pH 7.0 with 0. 1 M sodium phosphate
and used at 4°C. Paraffin wax serial sections, 4 or 8 ^m
thick, prepared by standard procedures, were stained by
Curtis's Ponceau S method for collagen, Ehrlich's hae-
matoxylin and eosin, Heidenhain's iron haematoxylin and
metanil yellow, or Mallory's trichrome stain. Histochem-
ical methods included an alcian blue, periodic acid-Schiff
(PAS) and orange G trichrome technique for glycoproteins
and mucosubstances, the mercury-bromphenol blue
method for proteins, Millon's and Sakaguchi's reactions
for tyrosine and arginine. Perls' method for ferrous and
ferric iron, and the ammonium hydroxide-alizarin method
for calcium (Pearse, 1972).
Polyphenol oxidase activity in the scales was detected
by a modification of Johri and Smyth's (1956) method;
formalin-fixed whole worms were treated with 0.1%
aqueous catechol (1,2 benzenediol) for 1 h, sectioned in
paraffin wax at 8 ^m and the sections dewaxed, mounted
in DPX, and examined using a deep blue filter transmit-
ting at 350-450 nm with peak transmission at 425 nm.
Controls were sections of untreated worms and whole
mounts of various proseriate and digenean worms showing
the vitellaria.
For ultrastructural studies, specimens were fixed for 3
h at 4°C in 3% glutaraldehyde buffered to pH 7.2 with
0.2 M phosphate, post-fixed for 1 h in buffered 1% osmium
tetroxide, embedded in Spurr's resin, and sectioned. Thin
sections, mounted on pioloform films carried on copper
slot grids and stained with uranyl acetate and lead citrate,
were examined in a JEOL 1200 EX transmission electron
microscope. Other sections, 1-2 ^m thick, were stained
with toluidine blue and studied with the light microscope.
The arrangement and general topography of the scales
were studied by light microscopy, using unstained for-
malin-fixed specimens cleared and mounted in DPX, and
by scanning electron microscopy of formalin-fixed worms
post-fixed in buffered 2% osmium tetroxide, processed by
standard procedures and examined in a Camscan Series
3SEM.
Results
Notodactylus handschini (Figs. 1, 2) is a broadly oval
temnocephalid, 1.0-1.5 mm X 0.75-1.0 mm at maturity,
with five anterior tentacles curled ventro-posteriorly when
at rest, a pair of eyes anteriorly, and a well-developed
sucker posteriorly. The entire dorsal epidermis is covered
by golden-brown scales bearing rich growths of epizoic
bacteria, cyanobacteria, diatoms, green algae, stalked cil-
iated protozoans, and sessile rotifers.
The scales lie in close-set rows but do not overlap; they
are up to 100 nm tall with rhombic bases 50-55 /urn by
15-20 /urn, whose long axes lie transversely to the worm's
longitudinal axis (Fig. 3). The great majority can be re-
ferred to a single basic form in which the rhombic cross-
section at the base continues upwards, decreasing in area
for some 1 5-20 nm before gradually becoming oval so
that the scales are essentially elongate conoids whose slen-
der tips curve inwards towards the worm's mid-line (Fig. 4).
Scales along the lateral body margins are always of this
shape but vary in height according to their position. Those
nearest the naked ventral epidermis are the smallest, rarely
more than 30-35 ^m tall, but the size increases across the
dorsal epidermis up to 55-65 ^m. In mature worms, the
more median scales may be reduced distally into squat
truncated conoids no more than 40-50 ^m tall (Fig. 6).
Their flat or slightly convex tops are covered in epizoic
growths of the same variety and abundance as those col-
onizing other surfaces of the scales, suggesting that loss
EPIDERMAL SCALES IN A TURBELLARIAN
119
of the curved tips is a normal consequence of aging. The
smaller, lateral scales bear only light growths, restricted
to their lower surfaces, supporting the conclusion that
they are younger than their more dorsal counterparts.
Most scales lying along the anterior and posterior body
margins are of the curved conoid type, but two or three
on each of the antero- and postero-lateral margins are
exceptionally tall, stout, and columnar, reaching 90-100
/urn in length (Fig. 6). Their cross-sectional shapes and
areas do not change along their length, and they remain,
in effect, tall rhombic prisms covered on all surfaces by
epizoic growths. Their tops, particularly, are prone to col-
onization by vorticellid ciliates. These columnar scales
are especially noticeable in living worms viewed by epi-
illumination, when they appear strongly iridescent.
Newly hatched N. handschini lack scales and are grey-
ish-white dorsally. Scale rudiments soon appear though
(Figs. 19, 20), and recognizable scales of the adult types
are present within three days. These are quickly colonized
by the characteristic assemblage of epizoites so that ju-
veniles four to five days old are indistinguishable externally
from adults, apart from their difference in size.
Retractile papillae, 100-150 yum by 30-40 nm when
extended, occur between the rows of scales sub-anteriorly
and posteriorly. They are simple outgrowths of the body
wall, are devoid of epizoites, and contain muscle fibers
continuous with the diagonal muscles of the general body
musculature (Fig. 6). They have no connection with the
scales and will not be described further here.
Histology and histochemistry of the scales, epidermis,
and rhabdilogen cells
Scales of all types and ages are strongly acidophilic,
staining deeply with eosin, orange G, and the acid fuchsin
and picric acid components of Mallory's and Curtis's
stains. They also stain strongly with toluidine blue, iron
haematoxylin, the mercury-bromphenol blue method for
proteins, Sakaguchi's method for arginine and the PAS
reaction. They stain only lightly with 1% aqueous alcian
blue prior to permanganate oxidation, but more deeply
subsequently, very lightly with Millon's reagent for ty-
rosine, and not at all with the Ponceau S component of
Curtis's stain for collagen and Perls' method for inorganic
iron. They react positively to the alizarin test for calcium,
especially basally; the reaction is strongest in formalin-
fixed scales, suggesting that the calcareous component is
susceptible to the acidic constituents of Bouin's fixative.
This was confirmed by treatment of formalin-fixed sec-
tions with 2% hydrochloric acid, which eliminated any
subsequent response to alizarin.
Iron haematoxylin staining followed by careful differ-
entiation in iron alum reveals darker staining bands in
the basal regions of mature scales, suggestive of growth
rings (Fig. 6).
The scales of formalin-fixed worms treated with 0.1%
catechol prior to sectioning at 8 ^m, showed a significant
darkening basally when compared with scales on un-
treated worms, indicating the presence of polyphenol ox-
idase or a similar quinone-tanning enzyme system. Be-
cause even young scales are golden-brown in color, such
darkening is difficult to discern with normal illumination,
but using a deep blue filter with peak transmission at 425
nm, the reactive zones showed greater absorption and were
clearly seen. Whole mounts and sections of various pro-
seriate and digenean flatworms, showing vitellaria or
eggshell-producing glands, acted as positive controls.
The plaque-like growths of epizoites on the scales pro-
vided useful controls for all these tests, with at least some
of the various organisms showing positive reactions to
one or another of them. Positive responses to Perls' test
for iron were particularly common.
The combination of reactions shown by the scales in-
dicates that they are glycoproteins tanned into a stable
physico-chemical form by a simple quinone-tanning sys-
tem. Their stability was demonstrated during the appli-
cation of the Millon's test for tyrosine when they survived
immersion in the reagent, containing 10% sulphuric acid,
for 5 min at 60°C — a procedure that destroyed all other
parts of the sections except the frustules of epizoic diatoms.
The epidermis beneath the scales is syncytial, as is that
covering the rest of the body. It is 5.0-5.5 ^m deep, with
strongly acidophilic cytoplasm, which stains deeply with
acid fuchsin, eosin, orange G, and mercury-bromphenol
blue. It reacts only weakly to PAS apart from the extreme
distal region, which gives a strong positive reaction (Fig.
5); this area appears as a striated border after iron hae-
matoxylin and is obviously the microvillar layer, which is
a dominant feature at the ultrastructural level (Figs. 8, 9).
Epidermal nuclei are infrequent but prominent, 6.5-
7.5 /*m by 4.5-5.0 nm, lying lengthwise in the syncytium
and with distinct, deeply staining chromatin. They may
cause the epidermis to bulge slightly inwards, but are never
insunk.
The epidermis rests on a thick fibrous basement mem-
brane, 7.0-8.0 nm deep, which stains strongly with Curtis's
Ponceau S method for collagen but only lightly with PAS.
The epidermis and basement membrane are traversed
by the slender necks of rhabdite-secreting gland cells
(rhabditogen cells), whose main bodies lie in the paren-
chyma below the dorsal subepidermal musculature (Figs.
5, 12). The rhabditogen cells occur throughout the dorsal
parenchyma but are commonest anteriorly, behind the
brain and above the pharynx, and posteriorly in the region
of the testes. They are ovoid to spherical, 40-50 /im in
diameter, with large nuclei and acidophilic cytoplasm
packed with rhabdites. The latter show all the staining
reactions given by the scales, including a positive response
to the alizarin test for calcium. Significantly, though, they
120
J. B. JENNINGS ET AL.
Figure 1. Notodactylus handschini, ventro-lateral aspect, showing the five tentacles (left), naked ventral
epidermis, posterior sucker, and portions of the latero-dorsal surface covered by epizoic microorganisms
growing on the epidermal scales. Some posterior scales (arrowed) bear only few epizoites. Scale bar = 200
jum.
Figure 2. Dorso-lateral aspect, showing the tentacles (right) and heavy growths of epizoites on the dorsal
and lateral surfaces. Scale as in Figure 1 .
Figure 3. Dorsal view of A', handschini photographed by dark-ground illumination after clearing and
mounting unstained in DPX. The focal plane is at the level of the rhombic bases of the scales; e. eyes. Scale
= 200 urn.
EPIDERMAL SCALES IN A TURBELLARIAN
121
show no reaction to PAS and alcian blue and are extremely
susceptible to mineral acids, rapidly disintegrating in the
10% sulphuric acid and 2% hydrochloric acid of Millon's
and Perls' reagents. Catechol has no effect on their ap-
pearance or staining properties. The rhabdites differ from
the scales, therefore, in their lack of carbohydrate and
polyphenol oxidase components and solubility in mineral
acids.
The cell necks of the rhabditogen cells follow a very
sinuous course through the parenchyma and musculature
to the epidermis and are almost impossible to trace in
their entirety, even in 8 ^m sections.
Ultrastructure of the scales
The scales are borne on rhombic epidermal plinths (Fig.
7), which have the same dimensions as the scales' bases.
The epidermal syncytium is not noticeably thickened to
form the plinths, but the plinth margins are produced
into shelf-like overhangs 2.0-2.5 ^m wide. These are sep-
arated from those of adjacent scales by spaces up to 5 /urn
wide. The epidermis dips downwards below the overhangs,
emphasising the plinth-like effect, but it is turned upwards
into a single fold equidistant between their tips. Each
plinth is thus surrounded by a shallow gutter, about half
as deep as the epidermis and separated from the adjacent
gutter by the epidermal fold. The scales do not move rel-
ative to each other during the worm's normal movements,
using the subepidermal musculature, and maintenance of
the scales' positions is presumably due to the hinge-like
action of the gutters and compensatory stretching of the
epidermal folds. We suggest, therefore, the term 'arthrodial
gutters' to describe these structures.
The scales are composed of ranks of uniform, closely
packed parallel tubes, 180-200 nm in diameter, and with
walls 40-45 nm thick (Figs. 7-11). The tubes run the
length of the scales, and the majority have no visible con-
tents; in sections cut obliquely to the scale's long axis,
they may have an apparently ordered basket-weave ar-
rangement, but examination of serial sections confirms
that this is an effect of the plane of sectioning. The base
of each tube encloses a single epidermal microvillus (Figs.
8, 9) but is not closely applied to it; a space 10-12 nm
wide remains between the tube- and microvillar walls and
is occupied by the glycocalyx. The tube bases, collectively
forming the base of the scale, do not rest directly on the
epidermal surface but appear to be supported some 80-
90 nm above it, presumably by their connection to the
microvilli via the glycocalyxes. This space was consistently
present, and of the same width, in all wax and resin sec-
tions examined and would not seem, therefore, to be a
shrinkage artefact. Both it and the tubes' lumina are pre-
sumably fluid-filled in life, with the fluid probably con-
tributing significantly to the scales' mechanical stability.
The tube walls are composed of electron-opaque gran-
ules, 0.5-1.0 nm in diameter, loosely assembled into
straight or slightly curved rod-shaped aggregates 20-30
nm by 8-10 nm (Figs. 9, 13, 15, 16). These tend to be
orientated with their long axes at 90° to the walls' long
axes. Most tubes lack visible contents, but a smaller num-
ber, 10-15% of the total, are twice the diameter of the
others and are packed throughout their length with a het-
erogeneous mixture of granules, similar to those of the
walls, and amorphous, less electron-opaque materials (Fig.
9). These larger tubes may each enclose a single micro-
villus basally. like the narrow tubes, or the microvilli may
be lost.
Tubes forming the central bulk of the scales are straight
and unbranched throughout their length. Those near the
scales' edges, however, curve outwards and often branch
dichotomously as they approach the edge (Figs. 10, 11).
The branches are always the same diameter as the parent
tube.
The scales are bounded by an unstructured layer 90-
100 nm thick, which is moderately electron-opaque and
formed from the disintegrating ends of the tubes. It is
most distinct and uniform along straight edges of the scales
near their bases (Fig. 10); it is less uniform on curved
Figure -4. Three conoid scales in vertical section. Two of the scales carry epizoic growths of various
microorganisms (arrowed); the middle scale shows the transition basally from rhombic to conoid shape.
Scale = 20 ^m.
Figure 5. Part of a sagittal section showing the basal regions of three scales (s), the strongly PAS-positive
distal border (microvillar layer) of the epidermis (arrowed), subepidermal musculature, and rhabditogen
cells (re) lying between blocks of diagonal muscles. Rhabdites in the cells are PAS-negative; their dark
appearance is due to their staining with orange G. Epizoites between the scales have stained deeply with
alcian blue and PAS. Section stained with alcian blue. PAS, and orange G. Scale = 20 ^m.
Figure 6. Longitudinal section through the anterior region showing a tall columnar scale (arrowed) with
bands, a papilla (p) whose muscle fibers extend into the parenchyma, and a truncated conoid scale (ts).
Section stained with iron haematoxylin and metanil yellow. Scale = 10 ^m.
Figure 7. Basal region of a conoid scale resting on its epidermal plinth, which is separated from adjacent
plinths by arthrodial gutters. Microvilli lining the gutters are smaller and less regular than those at the base
of the scale but bear long dense glycocalyxes (arrowed). Rhabditogen cell necks containing rhabdites are
passing through the epidermal plinth. Scale = 5.0 ^m.
122
J. B. JENNINGS ET AL.
edges, but here its origin from the walls is very obvious
(Fig. 1 1). The layers forming the upper surfaces of the
truncated scales occurring in the mid-dorsal region are of
this latter type, but are usually thicker, reaching 100-150
nm, and with very disorganized lower parts. The under-
lying tubes, unlike those at the sides of the scales, remain
straight and unbranched as they approach the surface,
suggesting that the level of the latter is determined by
attrition of a pre-existing curved tip.
Ultrastructitre of the dorsal epidermis in relation to the
scales
Dominant features of the syncytial dorsal epidermis
are the tall regular microvilli of the epidermal plinths be-
low the scales, shorter microvilli with long, dense glyco-
calyxes lining the arthrodial gutters, and the numerous
necks of parenchyma! rhabditogen and other cells which
pass through it to open at the bases of the scales.
The microvilli below the scales are evenly spaced col-
umns 1.25 nm by 0.08 ^m, without internal differentia-
tion, and with short rather granular glycocalyxes (Figs. 8,
9). Those lining the arthrodial gutters are smaller (only
0.2-0.25 pirn tall), but their glycocalyxes are much larger
and denser and appear as a thick fuzzy coat around the
microvilli and extending above them for 0.4-0.5 mm
(Fig. 7). They gradually become larger and more closely
spaced along the overhanging portions of the epidermal
plinths and grade into the upper surface types on the
shoulder regions where scale tubes begin to form around
them.
Most of the cell necks passing through the epidermis
are those of rhabditogen cells lying below the subepi-
dermal musculature in the dorsal parenchyma, whose
histological and histochemical properties are described
above. The cells' ultrastructure and method of rhabdite
production (Fig. 12) are the same as in other turbellar-
ians, including temnocephalids (see Smith et a/., 1982;
Williams and Ingerfeld, 1988), and need not be de-
scribed further here. Mature rhabdites leaving the cells
and migrating out to the epidermis along the cell necks
are elongate tapering rods, 1.50-1.75 ^m by 0.20-0.25
jum, elecron-opaque, and with a concentric lamellated
structure (Fig. 14). They change, however, as they reach
the distal epidermis; the internal lamellated structure
disappears, the electron-opacity may increase or become
much more heterogeneous, and they may become
curved (Figs. 13, 16, 17).
The rhabdites may be retained for a time in the distal
epidermis, apparently by terminal caps that seal off the
cell necks (Fig. 16), but are eventually discharged onto
the epidermal surface between the microvilli. On dis-
Figure 8. Part of the basal region of a conoid scale. Note the regular microvilli (mv) enclosed by the
bases of the scale tubes. Rhabditogen cell necks, some containing rhabdites. are visible in the syncytial
epidermis. Scale = 2.0 ^m.
Figure 9. Detail from the field seen in Figure 8, showing a large tube (left of centre) whose lumen is
packed with tube-wall building material. Grazing sections of walls of the commoner smaller tubes (arrowed)
show the rod-shaped aggregates of wall material. Scale = 500 nm
Figure 10. The uniformly structured layer bounding the basal region of a scale. Note the apparent
branching (arrowed) of some of the scale tubes. Scale = 500 nm.
Figure 11. The curved edge of the upper part of the scale. The boundary layer is not as well organized
as that shown in Figure 10. The apparent branching of scale tubes, with confluent lumina (arrowed) is clearly
seen; b, epizoic bacterium. Scale as in Figure 10.
EPIDERMAL SCALES IN A TURBELLARIAN
123
charge they disintegrate into the electron-opaque gran-
ules that form the principal components of the scale tube
walls (Fig. 15). The granules are, at first, rather disor-
ganized, but as they pass outwards between the microvilli,
they become orientated into the stacked rod-shaped ag-
gregates seen in the tube walls and in the lumina of the
larger tubes (Fig. 16). During their passage outward, the
aggregates themselves become automatically orientated
around the microvilli to form tubes, each of which is
separated from its microvillar template by the latter's
glycocalyx.
The cell necks of the rhabditogen cells are 400-450 nm
in diameter where they open onto the epidermal surface.
They are anchored here by inconspicuous zonulae ad-
haerentes lying immediately above prominent septate
desmosomes which encircle the necks to a depth of 450-
550 nm (Figs. 13, 15, 16). They are supported internally
by microtubules lying just below the cell membrane. Be-
low the desmosomes the necks may be separated from
the surrounding syncytium by apparent spaces, but these
are so inconsistent in their occurrence, shapes, and sizes
that they are probably shrinkage artefacts. Similar spaces
occur around the cell necks where they enter the epidermis
basally, and around the upward intrusions of the basement
membrane into the epidermis.
Cell necks delivering rhabdites to the epidermal surface
occur regularly throughout the epidermal plinths. Only
occasional ones occur in the portions overhanging the
arthrodial gutters, and these are curved as they divert from
the main plinth out into the overhangs.
Each scale is anchored to its epidermal plinth by cross-
striated fibrils that lie in cell necks opening onto the epi-
dermal surface beneath the corners of the scale's rhombic
base but inset from the overhanging portions. Each neck
contains a single fibril (Figs. 17, 18); in mature worms,
up to four such necks are present per corner, within a
roughly circular area 1.5-2.0 /*m in diameter. They are
not present in hatchlings possessing only rudimentary
scales but appear in juveniles, as the scales assume the
adult form, within 4-5 days of hatching.
The cell necks are similar to those delivering rhabdites
but are consistently larger, with neck diameters in the
range 550-600 nm and with the septate desmosomes ex-
tending down into the syncytium for 600-700 nm. Unlike
those of the rhabditogen cells, though, it was impossible
to trace them, with any certainty, beyond the subepider-
mal musculature and link them with a specific cell type
in the parenchyma. This was due to the absence of any
identifying structural or histochemical features within the
necks below the fibrils and the abundance of gland cell
types in the dorsal parenchyma.
The fibrils are cylindrical, 1.5-2.0 ^m long and 0.25-
0.30 /urn in diameter. They are provisionally identified as
collagen by virtue of their characteristic appearance, being
made up of regularly repeating units of dark and light
bands with a periodicity of 62.04 ± 0.36 nm (n = 78,
confidence limits 99%). This value was obtained from
pooled data gained by direct measurement of prints and
from scanning additional TEM negatives in a Fison's 'Vi-
tatron' densitometer, normally used for scanning electro-
phoresis gels. It was not possible to obtain histochemical
confirmation of their nature as the single fibrils could not
be located in paraffin wax sections.
Each fibril lies within the cytoplasmic sheath forming
the cell neck (Figs. 17, 18). Careful examination of serial
sections confirmed this intracellular location; the fibrils
do not lie extracellularly between parallel extensions of
the cytoplasm as is usual with collagen fibrils in other
animals. The cell necks are supported by microtubules,
and the cytoplasm generally contains two or three mito-
chondria closely adjacent to the fibrils (Fig. 17). In con-
trast, mitochondria were never seen in the cell necks of
rhabditogen cells.
The fibrils are inserted distally into the bases of the
wider scale tubes that are packed with rhabdite-derived
materials throughout their length (Fig. 18). They lose their
regular banded structure either just within the cell neck
opening or within a few nanometers of entering the scale
tube and the fibril ends become frayed and dispersed into
the tube contents. Proximally, the fibrils merge with the
cytoplasm of the cell necks; fixation and resolution were
not adequate for the details of fibril assembly to be seen.
Nothing was found to suggest that the fibrils are ciliary
rootlets or the bases of sensory structures. Cilia and ciliary
stubs, basal bodies, rootlets, and neuronal connections
were found in the groups of sensilla on the ventral surfaces
of the tentacles but nothing comparable was seen in as-
sociation with the fibrils.
The cytoplasm of the syncytium is very electron-opaque
and contains scattered mitochondria and profiles of cis-
ternae. Swollen cisternae often occur alongside the cell
necks (Fig. 13) but there are no indications of secretory
activities into the necks or microvilli, or on to the epi-
dermal surface.
Scale formation in young worms
Rhabditogen cells are dominant elements in the dorsal
parenchyma of worms fixed 6 h after hatching, and their
necks containing rhabdites are already present in the epi-
dermis and subepidermal tissues (Fig. 12). The epidermis
is syncytial and folded in a manner indicative of the future
positions of the epidermal plinths. Microvilli are well-
developed, especially on the upper surfaces of the folds,
and simple scale rudiments may be visible around these,
but most of the epidermis is naked.
Epidermal growth and folding continues and at about
1 2 h after hatching the future plinths and arthrodial gutters
124
J. B. JENNINGS ET AL.
Figure 12. Part of a section from a young N. handschini fixed 6 h after hatching. Rhabditogen cells (re)
packed with rhahdites are prominent in the dorsal parenchyma and rhabdites (arrowed) can be seen in
transit through subepidermal tissues and the epidermis. The epidermis is folded and microvilli are appearing.
Scale = 5.0 ^m.
Figure 13. A rhabdite (r) within a rhabditogen cell neck opening onto the epidermal surface between
the microvilli. A similar cell neck containing a microtubule (arrowed) but without a rhabdite lies nearby.
Note the granular aggregates, derived from discharged rhabdites, around the microvilli; c. swollen cisternae
in epidermis; sd, septate desmosome. Scale = 500 nm.
Figure 14. A rhabdite in transverse section within a cell neck, showing its lamellated structure. Scale
= 200 nm.
Figure 15. Remains of a discharged rhabdite lying between the bases of two microvilli. Scale = 200 nm.
Figure 16. Rod-shaped granular aggregates adding to tube bases between epidermal microvilli; ga, granular
aggregates; r, rhabdites; tc. terminal cap. Scale = 300 nm.
Figure 17. Part of a striated fibril (f) embedded in the cytoplasm of a cell neck. The cytoplasm contains
.•o mitochondria (m) and microtubules (arrowed): adjacent cell necks contain rhabdites (r). Scale = 500
•igure 18. A striated fibril in a cell neck with its distal end inserted into the base of a large tube; sd,
i "some. Scale = 300 nm.
are recognizable; the microvilli on the presumptive plinths
are longer than those in the gutters, scale rudiments are
present and many rhabdites are visible, passing through
the subepidermal musculature, basement membrane and
epidermis (Fig. 19).
Twenty-four hours after hatching, the basic shapes of
some epidermal plinths are established, with well-defined
gutters and overhangs (Fig. 20). A granular layer, up to
0.5 nm thick, is sometimes present at the level of the
microvilli but disappears in older worms; it is probably
EPIDERMAL SCALES IN A TURBELLAR1AN
125
rhahdite material poured from the epidermis but not yet
organized around the growing microvilli. Scale rudiments
at this stage are grey and soft but can be dissected from
the epidermis without losing their form, provided they
are not put under excessive pressure.
Subsequent development is very rapid, and in juveniles
3-4 days old the epidermis and scales are of the adult
type, with the scales' color changing from grey through
pale gold to golden brown. It is at this time, significantly,
that the basal regions of the scales first show a positive
catechol reaction for polyphenol oxidase.
Discussion
The temnocephalid Notodactylus handschini is unique
among known turbellarians in having its dorsal epidermis
covered by precisely formed and arranged permanent
scales. The only other reported occurrence of cuticular
structures in the Turbellaria is in the polyclad Enanlia
spinifera, which has epidermal spines along the body
margins (von Graff, 1 889). The spines form as a secretion
over an epidermal papilla, but the nature of the secretion
and the method of its stabilization are unknown.
Despite the unique nature of the scales in N. handschini,
their production and maintenance involve only precur-
sors, processes and structures occurring in one form or
another throughout the Turbellaria; the scales, therefore,
represent exploitation of existing features rather than the
evolution of entirely new ones.
The syncytial epidermis upon which the scales rest dif-
fers from that described in other temnocephalids (Wil-
liams, 1986, and references therein) in its lack of a distal
layer of vesicular epitheliosomes, the presence of striated
fibrils and its folding into epidermal plinths and arthrodial
gutters.
The rhabdites that contribute the bulk of the scale ma-
terial are of the lamellated type common elsewhere in the
temnocephalids (Williams, 1975, 1986; Williams and In-
gerfeld, 1988) and other turbellarians (Lentz, 1967;Bowen
and Ryder. 1974; Smith el al, 1982). In the temnoce-
phalids, they disintegrate after discharge onto the epider-
mal surface of the tentacles to form a thin surface film
which is stabilized by the microvilli (Williams, 1986). A
similar constant discharge, but over the entire body sur-
face, occurs in free-living turbellarians; the rhabdites hy-
drate and disintegrate to form a semi-fluid film, which is
thought to protect the otherwise naked ciliated epidermis
while still allowing ciliary activity (Jennings, 1957). The
protective film, composed of simple, unconjugated pro-
tein, is probably constantly renewed basally as it is eroded
or dissolved distally. This interpretation of the primary
function of rhabdites explains why they are produced in
such vast numbers, in most species, and constantly ex-
ported from their formative cells in the parenchyma into
and through the epidermis. A secondary function, but
still protective, is seen in polyclad turbellarians where both
cotyleans and acotyleans use them to form the large ge-
latinous masses in which the otherwise naked eggs are
embedded (Jennings, 1957).
In Notodactylus handschini, the protective role of the
rhabdites is taken much further by elaborating them into
permanent structures — the dorsal scales. Such scales, of
course, are incompatible with a ciliated epidermis and
ciliary locomotion but N. handschini, in common with
most other temnocephalids, has lost most of its external
ciliation and moves by muscular looping involving the
tentacles and posterior sucker.
The rhabdites in N. handschini are very similar in his-
tological and histochemical properties to those studied in
other turbellarians by Jennings (1957), Pedersen (1959),
Skaer (1961), and Bowen and Ryder (1974), and are
clearly homologous with these, a view confirmed by fur-
ther similarities in ultrastructure, method of secretion,
and mode of export to the epidermal surface within cy-
toplasmic strands of the formative cells. Their involve-
ment in scale formation, therefore, has not necessitated
any basic changes in these properties. The turbellarian
habit of continually discharging rhabdites through the
epidermis to maintain the protective surface film lends
itself readily to the formation and subsequent growth of
structures like the scales of N. handschini, provided that
the rhabdite-derived material can be stabilized. In N.
handschini, the stabilizing factor appears to be the com-
bination of the proteinaceous rhabdite material with a
carbohydrate moiety and the subsequent tanning of the
glycoprotein product by polyphenol oxidase. In view of
the histochemical properties of the scales, rhabdites, and
epidermis, the only possible source of this carbohydrate
would seem to be the glycocalyxes of the microvilli. The
polyphenol oxidase appears to be concentrated at the mi-
crovillar level, as could be expected, but its source is un-
known. Its occurrence, however, is not a novel feature as
it is commonly found throughout the Platyhelminthes as
a tanning agent in egg capsule production (von Brand,
1973).
The occurrence in the worm's mid-dorsal region of
scales that have lost their distal curved tops and become
reduced to truncated conoids, shows that, despite their
tanning, the scales are susceptible to erosion, perhaps by
water currents or the activities of their epizoites. The con-
stant addition of formative materials basally will com-
pensate for this, to some extent, just as the continual dis-
charge and disintegration of rhabdites in other species
maintains indefinitely the protective film over their cil-
iated surfaces.
The factors determining the curved conoid shape of the
majority of the scales remain unknown, along with the
reasons for the occurrence of the anomalous tall columnar
126
J. B. JENNINGS ET AL.
Figure 19. Part of the epidermis and subepidermal tissue of a young A', handschini fixed 12 h after
hatching. Rhabdites (arrowed) are passing through the subepidermal musculature and folded epidermis,
microvilli are well developed and scale rudiments are appearing; bm, basement membrane of epidermis.
Scale = 2.0 ^m.
Figure 20. An epidermal plinth and well-developed scale rudiment in a worm fixed 24 h after hatching.
Rhabdites (arrowed) can be seen below the scale rudiment. Scale = 2.0 fim.
scales on the anterior and posterior body margins. Anom-
alous, too, is the distal branching of some of the scale
tubes below the lateral surfaces of the scales (Figs. 10, 11).
The branches are all the same diameter as the other tubes,
suggesting, perhaps, that they result from fusion of the
distal regions of adjacent tubes rather than from true
branching. Alternatively, branching may occur as the
tubes form around the microvilli; occasional branched
microvilli do occur in young worms. If this is the case,
then the distribution of microvilli, including branched
ones, on the epidermal plinths is probably the decisive
factor in scale morphogenesis.
Microvilli and glycocalyxes are versatile structures put
to a variety of uses by animals. Examples are their roles
in membrane (contact) digestion in vertebrates (Ugolev,
1965), turbellarian adhesive systems (Tyler, 1976), cuticle
formation in oligochaetes and archiannelids (Potswald.
1971; Rieger and Rieger, 1976), cuticle attachment and
chaeta formation in annelids (Richards, 1978), and cuticle
attachment in pentastomid arthropods (Riley and Banaja,
1975). Their role in stabilizing rhabdite-derived films on
the tentacular surfaces of other temnocephalids (Williams,
1986) has already been mentioned; it is not surprising,
therefore, to find microvilli intimately involved in the
mechanics of scale formation in N. handschini, addition-
ally to the probable ctvjmical involvement of their gly-
cocalyxes. Their function as templates, around which the
glycoprotein becomes arranged to form the scale tubes,
is identical with the role of chaetoblast microvilli in the
formation of chaetae fiom a polymerizing chitin-protein
complex in annelids (Richards, 1978). In both instances.
the microvilli are long and extremely regular, in accord
with the long, regular tubes produced around them.
Attachment of the scales to their epidermal plinths is
probably another important function of the microvilli. It
is supplemented by the apparent collagen fibrils embedded
in the epidermis and inserted into the bases of some of
the larger tubes, which are probably strengthening pillars
because they are packed with material similar to that of
the tube walls.
If the striated fibrils are indeed collagen, then they are
the only features associated with the scales that are novel
epidermal structures; they have not been reported else-
where in the Turbellaria. But basement membranes are
usually collagenous (Burgeson, 1988), and that of TV.
handschini is probably no exception in view of its fibrous
nature and PAS-and Ponceau S-positive reactions. Thus
the fibril-secreting cells may well be homologous with
those secreting the basement membrane and other com-
ponents of the extracellular matrix.
Identification of the fibrils as collagen rests mainly upon
their size, appearance, and the periodicity of their banding.
The periodicity of 62.04 ± 0.36 nm lies well within the
range of 55.0-68.0 nm found in examples of collagens
taken from all major invertebrate groups from the Porifera
to Tunicata (Baccetti, 1985). In particular, it compares
with a value for Fasciola hepalica of 65.0 nm (Nordwig
and Hayduk, 1 969), which is the only other one available
from the Platyhelminthes. The fibrils' location beneath
the scales, their insertion into the bases of the strength-
ening pillars, their absence from hatchling worms pos-
sessing only rudimentary scales, and the absence of as-
EPIDERMAL SCALES IN A TURBELLARIAN
127
sociated basal bodies, recognizable ciliary stumps, and
neuronal elements (visible elsewhere in N. handschini) all
militate against an alternative interpretation of the fibrils
as ciliary rootlets. The association of mitochondria with
the fibrils (Fig. 17) might be regarded as supporting this
latter interpretation because mitochondria do occur
alongside the rootlets of monociliated sensory processes
in various turbellarians (Ehlers and Ehlers, 1977; Ferrero
and Bedini, 1989). In our opinion, though, this single fact
does not justify homologizing the fibrils with such rootlets,
especially in view of all the other evidence to the contrary.
The fibrils are reminiscient, in position and supposed
function if not in shape, of the U-shaped anchoring col-
lagen fibrils occurring in the epidermal-dermal and basal
lamina zones of vertebrates (Palade and Farquhar, 1965;
Bruns, 1969; Burgeson, 1988). According to Alberts et al.
(1989), collagen fibrils have the tensile strength of steel
so that although the fibrils in N. handschini are relatively
few in number, per scale, their concentration near the
corners of the scales' rhombic bases probably does provide
effective reinforcement of the scales' attachment to their
epidermal plinths via the microvilli.
The supposed collagen fibrils in N. handschini differ in
one outstanding respect from those occurring in other
animals and that is their indisputably intracellular location
over most of their length. There is ample and widely ac-
cepted evidence that collagen fibrils, in vertebrates at least,
form by self-assembly of their constituent molecules
within narrow extracellular compartments formed from
parallel but separate cytoplasmic extensions of the parent
fibroblasts (Birk and Trelstad, 1985, 1986; Burgeson,
1988). In A', handschini. the fibrils remain embedded in
the cytoplasmic sheaths forming the cell necks of their
formative cells and protrude from these only far enough
for insertion into the bases of the larger supporting tubes
of the scales. They pass through the distal neck regions,
which are encircled by the septate desmosomes locking
the necks into the epidermis; below this their accompa-
nying mitochondria confirm their intracellular position.
The occurrence of scales in N. handschini is probably
correlated with the worm's unusual life style. Other tem-
nocephalids are very active, but this species is remarkably
sedentary and remains for many days, perhaps for its en-
tire life span, at the same location on the edge of its host's
carapace (Cannon and Jennings, unpub. obs.). Mature
worms surround themselves with stockade-like circles of
their own egg capsules, some empty, others embryonated
or newly laid, and they remain quiescent within these for
long periods. They feed on small crustaceans, especially
ostracods, which settle on the eggs or nearby, by rapidly
extending the body, seizing the prey with their tentacles,
and swallowing it intact. This behavior, of course, protects
the eggs and can be construed as a form of brooding. The
scales' epizoites probably facilitate feeding by concealing
the worms from their potential prey; they also conceal
them from potential predators while the scales themselves
provide a protective shield over the body should an attack
occur. N. handschini is not known to have any particular
predators, but various other temnocepalids occur with it
on the same host (Cannon, 1991), and inter-specific pre-
dation is common in such communities on other cray-
fishes (Jennings, 1988).
Acknowledgments
This work was supported by a Royal Society travel grant
(SV/Australia 1990) to JBJ and an Australian Biological
Resources Survey grant (87/5909) to LRGC.
Literature Cited
Alberts, B., D. Bray, J. Lewis, M. Raff, K. Roberts, and J. D. Watson.
1989. Molecular Biology of the Cell. 2nd ed. Garland, New York.
1219pp.
Baccetti, B. 1985. Collagen and animal phylogeny. Pp. 29^17 in Biology
of Invertebrate and Lower Vertebrate Collagens. A. Bairati and R.
Garrone, eds. Plenum Press, New York.
Baer, J. G. 1945. Un Temnocephale nouveau. Temnocephala hand-
schini n. sp. de la Nouvellie Guinee. Rev. Suisse Zoo/. 52: 505-512.
Baer, J. G. 1953. Temnocephales. Zoological results of the Dutch New
Guinea Expedition 1939. No. 4. Zoo/. Meded. (Leiden) 32: 1 19-139.
Birk, D. E., and R. L. Trelstad. 1985. Fibroblasts create compartments
in the extracellular space where collagen polymerizes into fibrils and
fibrils associate into bundles. Ann. N.Y Acad. Sci. 460: 258-266.
Birk, D. E., and R. L. Trelstad. 1986. Extracellular compartments in
tendon morphogenesis: collagen fibril, bundle and macroaggregate
formation. J. Cell Biol 103: 231-240.
Blair, D., and J. B. Williams. 1987. A new fecampnd of the genus
Kronborgia (Platyhelminthes: Turbellana: Neorhabdocoela) parasitic
in the intertidal isopod Exosphaeroma obtusum (Dana) from New
Zealand. J. Nat. Hist. 21: 1 155-1 172.
Bowen, 1. D., and T. A. Ryder. 1974. The fine structure of the planarian
Polycelis tennis (lijima). III. The epidermis and external features.
Protoplasma 80: 381-392.
Brand, T. von. 1973. Biochemistry of Parasites, 2nd ed. Academic Press,
New York. 499 pp.
Bresciani, J., and M. Keie. 1970. On the ultrastructure of the epidermis
of the adult female of Kronborgia amphipodicola Christensen and
Kanneworff, 1964 (Turbellana: Neorhabdocoela). Ophelia 8: 209-
230.
Bruns, R. R. 1969. A symmetrical extracellular fibril. / Cell Biol. 42:
4 1 8-430.
Burgeson, R. E. 1988. New collagens. new concepts. Ann. Rev. Cell
Biol. 4:551-577.
Cannon, R. G. L. 1991. Temnocephalan symbionts of the freshwater
crayfish Cherax quadricarinatus from northern Australia. Hydro-
biologia ( in press).
Christensen, A. M., and B. Kanneworff. 1964. Kronborgia amphipod-
icola gen. et sp. nov., a dioecious turbellanan parasitizing ampeliscid
amphipods. Ophelia 1: 147-166.
Ehlers, L1., and B. Ehlers. 1977. Monociliary receptors in interstitial
Proseriata and Neorhabdocoela (Turbellana. Neophora). Zoomor-
phologit-86: 197-222.
Ferrero, E. A., and C. Bedini. 1989. Chemoreception in Turbellaria.
Exp. Biol. 48: 141-148.
Graff, L. von. 1889. Enantia spinifera, der Reprasentant einer neuen
Polvcladen-Familie. Nalurwiss. Vereines f. Steiermark 1889: 1-16.
128
J. B. JENNINGS ET AL.
Jagersten, G. 1942. Zur Kenntnis von Glanduloderma myzostomatis
n. gen. n. sp., einer eigentumlichen. in Myzostomiden Schmarotz-
enden Turbellarienform. Ark. Zoo/ 33A: 1-24.
Jennings, J. B. 1957. Studies on feeding, digestion, and food storage
in free-living flatworms (Platyhelminthes: Turbellaria). Biol. Bull
112:63-80.
Jennings, J. B. 1988. Nutrition and respiration in symbiotic Turbellaria.
Fortschr. Zool. 36: 1-13.
Jennings, J. B. 1989. Epidermal uptake of nutrients in an unusual
turbellarian parasitic in the starfish Coscinasterias calamaria in Tas-
manian waters. Biol. Bull. 176: 327-336.
Johri, L. N., and J. D. Smyth. 1956. A histochemical approach to the
study of helminth morphology. Parasitology 46: 107-117.
Lentz, T. E. 1967. Rhabdite formation in Planaria: the role of micro-
tubules. J. Ullraslr. Res. 17: 114-126.
Nordwig, A., and U. Hayduk. 1969. Invertebrate collagens: isolation.
characterisation and phylogenetic aspects. / Mol. Biol. 44: 161-172.
Palade, G. E., and M. G. Farquhar. 1965. A special fibril of the dermis.
J. Cell Biol. 27:215-224.
Pearse, A. G. E. 1972. Histochemistry: Theoretical and Applied, 3rd
ed. Churchill Livingstone, Edinburgh. 1518 pp.
Pedersen, K. J. 1959. Some features of the fine structure and histo-
chemistry of planarian subepidermal gland cells. Z. Zellforsch. 50:
121-142.
Potswald, H. E. 1971. A fine structural analysis of the epidermis and
cuticle of the oligochaete Aeolosoma bengalense Stephenson. J Mor-
phol. 135: 185-212.
Richards, K. S. 1978. Epidermis and cuticle. Pp. 33-61 in Physiology
of Annelids, P. J. Mill, ed. Academic Press, London.
Rieger, R. M., and G. E. Rieger. 1976. Fine structure of the archiannelid
cuticle and remarks on the evolution of the cuticle within the Spiralia.
Ada Zool. (Stockholm) 57: 53-68.
Riley, J., and A. A. Banaja. 1975. Some ultrastructural observations
on the integument of a pentastomid. Tissue Cell 7: 35-50.
Skaer, R. J. 1961. Some aspects of the cytology of Polycelis nigra. Q.
J. Microsc. Sci. 102: 295-317.
Smith, J., S. Tyler, M. B. Thomas, and R. M. Rieger. 1982. The mor-
phology of turbellanan rhabdites: phylogenetic implications. Trans.
Am. Microsc. Soc. 101: 209-228.
Tyler, S. 1976. Comparative ultrastructure of adhesive systems in the
Turbellaria. Zoomorphologie 84: 1-76.
Tyler, S. 1984. Turbellanan Platyhelminths. Pp. 112-131 in Biology
oflhe Integument, I '«/. / Invertebrates, J. Bereiter-Hahn, A. G. Ma-
tolsky, and K. S. Richards, eds. Springer- Verlag, Berlin.
Ugolev, A. 1965. Membrane (Contact) digestion. Physiol. Re\' 45: 555-
595.
Williams, J. B. 1975. Studies on the epidermis of Temnocephala
I. Ultrastructure of the epidermis of Temnocephala novae-^ealandiae.
Ausl. J. Zool. 23: 321-331.
Williams, J. B. 1980. Studies on the epidermis of Temnocephala
V. Further observations on the ultrastructure of the epidermis of
Temnocephala novae-zealandiae. including notes on the glycocalyx.
Aust. J. Zool. 28: 43-57.
Williams, J. B. 1986. Phylogenetic relationships of the Temnoce-
phaloidea (Platyhelminthes). Hydrobiologia 132: 59-67.
Williams, J. B., and M. Ingerfeld. 1988. Cells in the parenchyma of
Temnocephala: rhabdite secreting cells of Temnocephala novae-zeal-
andiae (Temnocephalidae: Platyhelminthes). Int. / Parasiiol. 18:
651-659.
Reference: Biol. Bull 182: 129-134. (February, 1992)
Red Blood Cell Oxygen Binding
in Capitellid Polychaetes
CHARLOTTE P. MANGUM1, JAMES M. COLACINO2, AND JUDITH P. GRASSLE3'*
^Department of Biology, College of William and Mary, Williamsburg, llrginia 23185,
2Department of Biological Sciences, Clemson University, Clemson. South Carolina 29634-3581,
and ^Marine Biological Laboratory, Woods Hole, Massachusetts 02543
Abstract. The oxygen equilibrium properties of red
blood cells that circulate in the coelomic cavities of 10
morphologically similar capitellid polychaetes are gen-
erally species specific. Appreciable differences in oxygen
affinity distinguish most of the species and, in several in-
stances, cooperativity differs as well. The range of oxygen
affinities is far greater than in the other closely related
polychaetes examined to date. We suggest that the differ-
ences may prove to be adaptations to thermal properties
of the environment, body size, or both.
Introduction
The polychaete Capitella capitata (Fabricius) is now
known to be a number of morphologically similar species
that are distinguished by a variety of genetic, develop-
mental and reproductive features (Grassle and Grassle,
1976; Eckelbarger and Grassle, 1987a,b; Grassle et a!..
1987). These features clearly indicate that the cryptic spe-
cies are in fact reproductively isolated from one another.
Like other members of the family Capitellidae, these spe-
cies contain typical annelid red blood cells (RBCs) that
circulate in the coelomic cavity.
The respiratory properties of the annelid RBC hemo-
globins (Hbs) are known primarily from investigations of
the families Glyceridae and Terebellidae (reviewed by
Mangum, 1992a). Although no direct comparisons of the
Hbs of congeneric or otherwise closely related species have
been made, they appear to be fairly similar within each
family. These Hbs have high (terebellids) to moderate
Received 9 July 1991: accepted 30 September 1991.
* Present address: Department of Marine and Coastal Sciences, Rutgers
University, P. O. Box 231. New Brunswick, New Jersey 08903.
(glycerids) O: affinities, and little or no cooperativity and
pH dependence. Although differences between the O:
binding properties of RBCs and those of purified Hbs have
been reported on a few occasions, thorough reinvestigation
has indicated that the annelid RBC Hbs are not sensitive
to intracellular effectors (Mangum, 1992a). The three ad-
ditional annelid families with RBCs have received scant
attention.
Most members of the genus Capitella are very small
animals (e.g., as small as 0.28 mg in the present sample),
and a broad survey of the O2 binding properties of their
Hbs with the techniques presently in wide use, even those
that require only several drops of material would be dif-
ficult. The recent development of a microspectrophotom-
eter specifically built for optically heterogeneous prepa-
rations (Colacino and Kraus, 1984), however, has made
it possible to determine the O2 binding of a single red
blood cell (Mangum et ai, 1989). In the present investi-
gation, we have taken advantage of this capability to com-
pare the respiratory properties of the RBCs of six of the
species previously regarded as Capitella capitata, and four
belonging to the closely related genus Capitomastus. We
have also attempted to relate our findings to O2 uptake.
Materials and Methods
The animals were obtained from laboratory cultures,
either strains that had been maintained continuously for
long periods or collected recently and cultured until they
produced brood tubes and viable larvae. With one excep-
tion, the species designations, collection sites, and culture
conditions were described by Grassle et al. (1987). Cap-
itomastus VOZ was collected in the thermal effluent af-
fected vicinity of a power plant (Millstone) in Connecticut,
and cultured at 20°C in 31.8 %» seawater.
129
130
C. P. MANGUM ET AL.
O: binding
Several individuals of a species were immersed in buf-
fered seawater on the side of a shallow concavity in a
depression slide and their body walls slitted, with iridec-
tomy scissors in the case of the larger animals and with
fine pins in the case of the smaller ones. On the few oc-
casions when the gut was ruptured, the material was dis-
carded. Otherwise, the RBCs were allowed to gravitate in
the central depression free of other body parts and were
then transferred to the experimental chamber. This
chamber, illustrated by Colacino and Kraus (1984), holds
a thin layer of RBCs between two polypropylene mem-
branes (25 nm thickness). A humidified mixture of Re-
search Grade N2 and O2, prepared in the present inves-
tigation by a gas mixing flowmeter (Cameron Instruments
Co.), was passed on either side of the polypropylene sand-
wich, which was held in a water-cooled brass slide. RBCs
were illuminated through glass windows in the brass. In
the present investigation, the diode array microspectro-
photometer described earlier (Mangum el a/., 1989) was
used exclusively.
With this apparatus, the entire visible spectrum in the
range 380-650 nm is collected. We saw no sign of the
spectral changes, denaturation, or precipitation encoun-
tered by Wells and Warren (1975) in their investigation
of capitellid RBCs, possibly because our RBCs were never
exposed to air. Initially, however, several experiments were
aborted by cell lysis; the problem was eliminated by first
soaking the polypropylene membranes in shaking sea-
water.
Data analysis
Hill plots of the data were analyzed by each of two
procedures. First, each RBC of each species was treated
as a unique member of the population. The constants P50
(PO2 at 50% oxygenation, or O2 affinity) and n50 (slope
of a Hill plot of O: binding at P5(l, or cooperativity) were
obtained from regression lines describing the data for in-
termediate oxygenation states. The mean values for each
species were compared by Student's / test. Second, the
RBCs of a particular species were treated as members of
a homogeneous population, and a single regression line
was constructed for all data in the range 10-90% HbO2.
Ninety-five percent confidence intervals around the
regression lines and their slopes are reported below as the
error.
O: uptake
Oxygen uptake of 2-20 animals, depending on body
size (see Results for size range), was determined by re-
cording the decline in PO2 in a closed container, using a
polarographic electrode (Strathkelvin Instrument Co.).
The volume of the chamber was 0.4-0.75 ml; stirring was
provided from the bottom with a magnetic bar which was
isolated from the animals by a disk of stainless steel mesh.
Temperature change was controlled to less than 0. 1 °C.
Results
Like Pals and Pauptit (1979), we noted that in vivo the
RBCs circulate in clumps. While many kinds of RBCs
aggregate in vitro when the medium is unstirred, we are
unaware of additional reports indicating that the phe-
nomenon is natural. We also noted distinctive and non-
overlapping differences in RBC diameter in several of the
species, suggesting that a hematological investigation
would be of interest.
RBC O2 binding
Oxygen affinity of the 10 species examined here ranges
from high (2.8 mmHg) to moderately low (11.1 mmHg).
Apparent cooperativity ranges from very little (1.6) to
pronounced (3. 1 ). While the RBCs of a few of the species
clearly showed increases in apparent cooperativity at high
oxygenation states (e.g., the increase in slope of the line
in Fig. 1 ), most did not. We did not investigate pH de-
pendence.
When treated as unique entities, the RBCs of a species
generally have distinctive O2 affinities, the mean values
of which differ significantly (P < .05) from those of most
of the other species. Several of the cooperativity values
also differ significantly from the others, although less fre-
quently (Tables I, II).
When treated as members of a homogeneous popula-
tion, the RBCs of a particular species have O2 affinity
>-
I
1.0-
0.1
8
o
o
o
1.5 10.0
P02 (mmHg)
Figure 1. Hill plot (log(y/l - y) vs. log PO,) ofO2 binding of RBCs
from Capitella MB/sl. 20°C, pH 7.4.
O, BINDING IN CAPITELLIDS
131
Table I
O: equilibrium constants" for RBCs of capitellid polychaetes
Species
P50 (mmHg)
"so
Capitella
sp. 1
6.88 ± 0.97
2.89 ± 0.82
7
sp. IA
7.99 ±0.61
2.36 ± 0.38
17
sp. II
5.47 ± 0.43
3.27 ± 0.55
10
sp. IMA
5.76 ± 1.17
2.61 ±0.84
6
MB/sl
8.27 ±0.82
1.68 ±0.46
6
ORL
11.3 ±0.9
3.44 ± 0.55
11
Capitomaslus
NCS
9.74 ± 0.93
3.62 ±0.91
11
TAR
3.85 ± 0.20
2. 76 ±0.33
14
TRIN
3.09 ± 0.79
2.56 ± 1 .03
9
VOZ
4.32 ± 0.93
3.03 ± 1.05
6
a Mean + S.E., 20°C, pH 7.4.
values that invariably differ significantly from all others
(Table III). Again, significant differences in apparent
cooperativity are found, but far less frequently (Table III).
O2 uptake
Oxygen uptake measurements were performed suc-
cessfully on five of the ten species. Three of the remaining
species were unavailable at the time. Individuals of Cap-
itella species IIIA and Capitomastus NCS, though avail-
able, were so small (<1 mg wet wt) and so geonegative
that they were able to (and did) crawl out of the chamber
through a small groove in the electrode housing that serves
as the route of gas escape during sealing.
As expected, the weight-specific rate of O: uptake (VO2)
varies inversely with body size (Fig. 2). VO2 was influenced
by ambient PO2 (PiO2) even less in the present investi-
gation than reported earlier (Mangum and Van Winkle,
1973), in all likelihood due to smaller body size.
Discussion
While the two analytical procedures might seem to yield
occasionally contrary results for significant differences in
O2 equilibrium constants, we suggest that they essentially
agree from a physiological point of view. Although the
analysis reported in Table III requires fewer assumptions,
several of the significant differences are trivial. For ex-
ample, the RBC O2 affinities of Capitella sp. II and IIIA
are essentially the same, as are those ofCapitomatsus VOZ
and Capitomastus TAR. The difference between Capitella
sp. I and I A is also small. Of possible relevance, two of
these species pairs are known to share the same chro-
mosome numbers. In Capitella sp. II and Ilia, 2n = 26;
Table II
Significance of(P < .05) differences in mean values oj O, equilibrium constants in Table I
A. P50
Capitella
sp. 1
sp. IA
sp. II
sp. IIIA
MB/sl
ORL
Capitomastus
NCS
TAR
TRIN
B. nso
Capitella
sp. I
sp. IA
sp. II
sp. IIIA
MB/sl
ORL
Capitomastus
NCS
TAR
TRIN
sp. IA
sp. II
+
+
Capitella
sp. IIIA
Capitomastus
MB/sl
ORL
NCS
TAR
TRIN
VOZ
132
C. P. MANGUM ET AL.
Table III
Oxygen equilibrium constants and their 95",: confidence limits
as estimated by regression amih i
Species
Pso(mmHg) L,
"so
Capitella
sp. I
7.05
6.65
7.45
2.19
1.27
3.11
sp. IA
7.75
7.68
7.82
2.24
2.05
2.43
sp. II
5.36
5.20
5.48
2.90
2.41
3.39
sp. IIIA
5.79
5.56
5.97
2.77
2.11
3.58
MB/sl
8.35
8.19
8.48
1.66
1.22
2.00
ORL
11.1
10.9
11.3
3.13
2.73
3.47
Capitomastus
NCS
9.54
9.37
9.71
2.84
2.32
3.36
TAR
3.91
3.86
3.97
2.59
2.28
2.90
TRIN
2.76
2.68
2.84
2.02
1.65
2.39
VOZ
4.18
3.97
4.42
2.98
2.12
3.84
in Capitella sp. I and la, 2n = 20 (Grassle et ai, 1989).
The chromosome number in Capitomastus VOZ is not
known.
More importantly, most of the interspecific differences
in O2 affinity are large and physiologically meaningful.
The values can be grouped as follows ( 1 ) Capitomastus
TRIN, (2) Capitomatsus VOZ and Capitomastus TAR,
(3) Capitella sp. II and IIIA, (4) Capitella sp. I, I A and
MB/sl, (5) Capitomastus NCS, and (6) Capitella ORL.
To emphasize the magnitude of the interspecific dif-
ferences, the range found in the present investigation is
twice as great as that known in four species in the genus
Glycera (Weber, 1978), which have never been regarded
as cryptic. This comparison suffers from the unequal
sample sizes. However, if the sample is enlarged and
broadened to include all nine annelid and echiuroid RBC
Hbs examined in the past two decades at a common tem-
perature of 20°C (Mangum, 1992a), then the range re-
ported here is still somewhat greater, despite the far nar-
rower taxonomic scope. We suggest that the large range
of O2 affinities is due to the similarly large geographic
range of our sample (see below). In contrast, the values
in the literature do not represent tropical or otherwise
warm waters.
We describe the 'cooperativity' of capitellid RBCs as
apparent because many RBCs containing only monomeric
and, when extracted and purified, non-cooperative Hbs
exhibit Hill slopes significantly greater than 1 at high ox-
ygenation stai.'s U:g., Mangum, 1976). This cellular level
phenomenon, although very common, is not understood.
The only information on the molecular weight of a cap-
itellid Hb other than the anomalously low value of 12 kD
for Heteromastusfiliformis (Pals and Pauptit, 1979) is the
figure of 34.5-36.4 kD for Notomastus latericeus Hb re-
ported by Svedberg and Eriksson-Quensel (1934), which
suggests a dimer. Bivalve dimers exhibit significant coop-
erativity, although not as much as found in several of the
RBCs examined here. The nature of the apparent coop-
erativity exhibited by capitellid RBCs will remain unclear
at least until their Hbs have been investigated in more
detail.
Environmental correlates
Because the species in our sample have been recognized
as such only recently, their full ecological and geographic
ranges are essentially unknown. All of our material was
collected from fine sediments in the intertidal or shallow
subtidal zones. We know of no factors, behavioral or eco-
logical, that would differentiate them according to ambient
PO2. Their considerable morphological similarity also
precludes the likelihood of physiological differences arising
from the design of gas exchange and transport systems.
We tentatively suggest that at least some of the differences
observed here may prove to be related to thermal char-
acteristics of the environment. If the P50 values are ar-
ranged in series (Table IV), the highest O2 affinities are
found in species living in the warmest waters, and the
lowest in species inhabiting colder waters. This relation-
ship is far more clearly characteristic of a variety of other
O2 carriers (Mangum, 1992a,b). It results in more similar
O2 affinities at the natural temperatures than would exist
in the absence of the adaptation.
Body size
None of the animals examined here exceeded 1 mm in
diameter, and thus O: uptake would not be expected to
be limited by the diffusion distance until the driving PO2
gradient reaches quite low levels. The rate of O: depletion
began to depart from linearity only below 50 mmHg (e.g.,
Fig. 3) and continued to do so until O2 uptake ceased.
200
E
o>
Body Wet Wt (mg)
Figure 2. The relationship between O2 uptake and body size in five
capitellids at 20°C. The smallest animals were juveniles of Ciipi/c/lii sp.
I; the remainder were adults.
O, BINDING IN CAPITELLIDS
133
Table IV
i affinity in relation to known geographic distribution
20
P50 (mmHg)
at pH 7.4
Species
Distribution
2.76
Capitomastus TRIN
Trinidad
3.91
Capilomastus TAR
North Carolina
4.18
Capitomastus VOZ
Connecticut (thermal effluent)
5.36
Capilella sp. II
Marseille, Boston Harbor,
and Cape Cod
5.79
Capilella sp. IIIA
Boston Harbor and Cape Cod
7.05
Capitella sp. I
Temperate zone cosmopolitan
7.75
Capitella sp. IA
Boston Harbor to New
Bedford
8.35
Capitella MBsl
Mission Bay, California
9.54
Capitomastus NCS
Uncertain, perhaps Mission
Bay, California
11.1
Capitella ORL
Cape Cod
The non-linear portion of the records (Fig. 3) represents
a combination of (1) Hb deoxygenation and its contri-
bution to the O2 being consumed and (2) the diffusion-
limited component of O: uptake. This portion can be
bisected to provide a numerical estimate of the ambient
PO: (PiO:) at which these two processes occur, best ex-
pressed as the half constant [e.g.. the intersect of the dashed
line and the trace of PO: in Fig. 3. The graphical procedure
is essentially the same as that detailed by Colman and
Longmuir (1963)]. This constant, which would be ex-
pected to be strongly influenced by both O: affinity and
the magnitude of the diffusion limitation, is in fact highly
correlated with body size (r = 0.944; P < .05) and, even
more strongly, with P50 (r = 0.977; P < .01 ) (Fig. 4). The
relationship between O2 affinity, environmental temper-
ature, and body size is actually syllogistic, with larger body
size being found at higher latitudes and smaller animals
found at lower latitudes. An interrelationship between
PO2, body size, and growth rate has been recently reported
in Capitella sp. I by Forbes and Lopez (1990).
P02
Figure 3. The decline in PCK with time due to O2 uptake by two
individuals (mean body wet wt. 39.4 mg) of Capitella ORL at 20°C. The
shaded area represents the portion of Oi uptake that is influenced by the
release of Hb bound CK and by a limiting diffusion distance. The dashed
line identifies the PiO: of half of the non-linear portion of O: uptake.
15
O
in
o
El
10
A
(ORL)
• 00
5 - m OWN)
5 10 15 20 25
Body Wet Wt (mg)
20
o
m
o
Q_
15
10
B
• (ORL)
(I)
10
15
P,n (mmHg)
so
Figure 4. The relationship between the PiO, of half of the non-linear
component of O: uptake (PiO:50) to body size (panel A) and O2 affinity
(panel B).
Most important in the present context, the respiratory
properties of the RBCs are distinctive of most of the species
examined here, for whatever adaptive reason. In view of
the conservatism of the annelid RBC Hbs investigated
previously (Weber, 1978; Mangum. 1992a), this was not
necessarily an expected finding.
Acknowledgments
Supported by NSF DCB 88-16172 (Physiological Pro-
cesses) to CPM, DMD 86-00614 (Biological Instrumen-
134
C. P. MANGUM ET AL.
tation) to JMC, and BSR 86-00648 (Systematic Biology)
to JPG.
Literature Cited
Colacino, J. M., and D. W. Kraus. 1984. Hemoglobin-containing cells
of Neodasys (Gastrotricha, Chenotida). II. Respiratory significance.
Comp. Bioclwm. Physiol. 79A: 363-370.
Colman, C, and I. S. Longmuir. 1963. A new method for registration
of oxy-hemoglobin dissociation curves. J. Appl. Physiol- 18: 420-
423.
Eckelbarger, K. J., and J. P. Grassle. 1987a. Spermatogenesis, sperm
storage and comparative sperm morphology in nine species of Cap-
itella, Capitomastiis and Capitellides (Polychaeta: Capitellidae). Mar.
Bio/. 95:415-429.
Eckelbarger, K. J., and J. P. Grassle. 1987b. Interspecific variation in
genital spine, sperm, and larval morphology in six sibling species of
Capitella. Bull. Biol. Soc. \l'ash. 7: 62-76.
Forbes, T. L., and G. R. Lopez. 1990. The effect of food concentration.
body size, and environmental oxygen tension on the growth of the
deposit feeding polychaete, Capitella species 1. Limnol. Oceanogr.
35: 1535-1544.
Grassle, J. P., and J. F. Grassle. 1976. Sibling species in the marine
pollution indicator Capitella (Polychaeta). Science 192: 567-569.
Grassle, J. P., C. E. Gelfman, and S. W. Mills. 1987. Karyotypes of
Capiiella sibling species and of several species in the related genera
Capitellides and Capitomastiis (Polychaeta). Bull Biol Soc. Wash.
7: 77-88.
Mangum, C. P. 1976. Primitive respiratory adaptations. Pp. 191-278
in Adaptation to Environment: Physiology of Marine Animals, R. C.
Newell, ed. Butterworth's & Co., Ltd.. London.
Mangum, C. P. 1992a. Respiratory functions of the red blood cell he-
moglobins of six animal phyla. In Blood and Tissue Oxygen Carriers,
C. P. Mangum. ed. Adv. Comp. Env. Physiol., Springer-Verlag, Hei-
delberg (in press).
Mangum, C. P. 1992b. Respiratory function of the molluscan hemo-
cyanins. In Blood and Tissue Oxygen Carriers, C. P. Mangum, ed.
Adv. Comp. Env. Physiol., Springer-Verlag, Heidelberg (in press).
Mangum, C. P., and W. Van Winkle. 1973. Responses of aquatic in-
vertebrates to declining oxygen conditions. Am. Zool. 13: 529-541.
Mangum, C. P., J. M. Colacino, and T. L. Vandergon. 1989. Oxygen
binding of single red blood cells of the annelid bloodworm Glycera
dibranchiata. J. Exp. Zool. 249: 144-149.
Pals, G., and E. Pauptit. 1979. Oxygen binding properties of the coe-
lomic haemoglobin of the polychaete Hete romastus filiformis related
with some environmental factors. Neth. J Sea Res 13: 581-592.
Svedberg, T., and I-B. Eriksson-Quensel. 1934. The molecular weight
of erythrocruonn. II. J. Am. Chem. Soc. 56: 1700-1706.
Weber, R. E. 1978. Respiratory pigments. Pp 393-446 in Physiology
of Annelids. P. J. Mill, ed. Academic Press, London.
Wells, R. M. G., and L. M. Warren. 1 975. The function of the cellular
haemoglobins in Capiiella capitata (Fabricius) and Nolomastus la-
tericeus Sars (Capitellidae: Polychaeta). Comp. Biochem. Physiol 51 A:
737-740.
Reference: Biol. Bull. 182: 135-144. (February, 1992)
Visual Cells and Pigments in a Demersal Fish,
the Black Sea Bass (Centropristis striata)
K. V. SINGARAJAH* AND F. I. HAROSI
Laboratory of Sensory Physiology. Marine Biological Laboratory, Woods Hole,
Massachusetts 02543, and Department of Physiology, Boston University
School of Medicine, Boston, Massachusetts 021 IS
Abstract. Using a single-beam, wavelength-scanning,
dichroic microspectrophotometer, we measured absolute
absorbance. bleaching difference absorbance, and linear
dichroism spectra from isolated retinal receptors of the
black sea bass, Centropristis striata. We determined,
among other properties, the wavelength of peak «-band
absorbance (Xmax) of the pigment of the receptor cells.
Out of well over 100 recordings, we found only 3 spectral
types of visual pigment. The shortest-wavelength-absorb-
ing type (Xmax = 463 ± 2 nm) was present only in single
cones. Both members of the double cones contained the
longest-wavelength-absorbing pigment of the three, with
Xma, = 527 ± 5 nm. Rods were found to bear a typical
rhodopsin, with Xmax = 498 ± 2 nm. Thus, the retina of
this predatory demersal fish appears to use a set of three
closely spaced visual pigments, with Xmax clustering about
500 ± 30 nm. This remarkable feature is discussed in
relation to photic conditions in the habitat.
Introduction
Because the function of an eye is to detect light from
the environment, visual systems must have evolved in
harmony with the prevailing photic conditions. An ad-
aptation of certain aspects of eye structure to an animal
habit was clearly recognized by Schultze (1866, 1867).
Being an exceptionally keen observer, he noted a corre-
lation between the preponderance of retinal rods in the
eyes of nocturnal animals, and the occurrence of numer-
ous cones in the retinas of diurnal animals. He rightly
reasoned that there is no color perception at night, that
Received 3 June 1991: accepted 30 September 1991.
* Permanent address: Laboratory1 of Marine Biology & Comparative
Physiology. DSE-CCEN. Federal University of Paraiba. Joao Pessoa—
58059, Pb.. Brazil.
nocturnal animals are therefore adapted to dim (scotopic),
black-and-white conditions, and that their vision is pri-
marily mediated by rods. Diurnal animals, on the other
hand, are mainly exposed to bright (photopic) conditions,
during which color sensation is most acute, and cones
must therefore be the primary mediators of color vision.
Thus, Schultze's observations provided the basis for what
is now known as the Duplicity Theory, which encom-
passes some of the most basic features of vertebrate vision.
A second link, between vision and habitat, was realized
by Clarke (1936). He knew that the peak transmittance
of pure water is in the blue part of the spectrum; and the
deeper the water column to be penetrated, the more
"squeezed" is the daylight spectrum about the blue peak.
On the basis of this understanding, he suggested "the pos-
sibility of a shift in sensitivity of the eye of a deep water
fish toward the blue end of the spectrum." Indeed, not
only do fishes of the deep sea have retinas with numerous
long rods, but their photosensitive pigments (rhodopsins)
also have blue-shifted peak absorbance (Xmax) values to
match the dominant wavelength of the scarce quanta
available to them (Denton and Warren, 1957; Munz,
1958). From these observations stemmed the Sensitivity
Hypothesis, which proposes an adaptation to the photic
environment by the selection of visual pigments that ren-
der the animal's eye most sensitive to the ambient illu-
mination.
To account for the apparent mismatch between the
underwater light and the Xmax of the visual pigments in a
number of animals, other than deep sea fish, the Contrast
Hypothesis was proposed (Lythgoe, 1972). The merit of
the underlying idea is the recognition that the differential
scattering and selective absorption of underwater light
may cause an object viewed against its background space-
light to be more visible with offset visual pigments than
with matched ones. Thus, brightness contrast detection.
135
136
K. V. SINGARAJAH AND F. I. HAROSI
in addition to sensitivity, is an important aspect of vision.
So is color perception, in which Xmax variation of the visual
pigment can play a major role. The question, at present,
is not so much about the relative importance of these
attributes in general, but rather their specific contribution
to the evolution of the eye.
Although we still lack a unified theory of visual func-
tion, a refinement in the Sensitivity Hypothesis goes a
long way toward explaining several vision-related phe-
nomena. This refinement resulted from a series of exten-
sive investigations (Munz and McFarland, 1973; Mc-
Farland and Munz, 1975a, b, c) in which not only the
visual pigments were determined, but also the spectral
radiance of natural light under many different environ-
mental conditions was measured. It was found that the
twilight spectrum was generally well matched by the sco-
topic pigments. Thus, the emerging concept was that the
spectral location of scotopic visual pigments have been
selected to enhance photosensitivity at twilight, for it is
during this period that visual behavior is critical to survival
(McFarland and Munz. 1975c).
Investigation ofihephotopic visual pigments has barely
begun. Although surveys by Loew and Lythgoe (1978)
and Levine and MacNichol ( 1979) are important contri-
butions, providing correlations between habitat and the
photopic pigments (see also Lythgoe. 1984), the devel-
opment of this area of research is still in its infancy.
For the present study, we chose a day-active predatory
fish that inhabits a marine environment with a fairly well
delineated photic habitat. By the use of microspectropho-
tometry. we set out to determine the light absorbing prop-
erties of its visual pigments throughout the visible and
near ultraviolet spectrum. We also made an attempt to
estimate the spectral light available in the fish's environ-
ment. We report here our findings concerning the spec-
troscopic relationship between visual pigments and the
ambient irradiance.
Materials and Methods
Experimental animals
Specimens were obtained through the courtesy of the
US National Marine Fisheries Service Aquarium at
Woods Hole. The fish we used were small specimens that
had been trapped in Woods Hole Harbor late in the sum-
mer of ' f>90 (about 6 months prior to their use) and
maintain, natural seawater aquaria at about 18°C
(65°F). The ' ere regularly fed with brine shrimp,
clams, and squid, and were kept under artificial illumi-
nation (8 h on/ 16 h off) with some natural light filtering
into the building over their tanks.
Preparation
Before use, each fish was dark-adapted for at least I h,
and anesthetized with aerated seawater containing tricaine
methanesulfonate (MS222, Sigma Chemical Company)
at a concentration of 0.25 g per liter. Enucleation of the
eye was performed under dim red light. The cornea, iris,
and lens were removed by a cut with a sharp razor blade,
transverse to the anatomical axis, about 1 mm behind the
equatorial circumference of the eye, and the entire eyecup
was quickly transferred to cold saline solution. With the
aid of a modified, low-power dissecting microscope
equipped with infrared illumination and an image con-
verter, the retina was eased away from the back of the eye
while the eyecup still submerged. From the whole retina,
small pieces (about 1 mm2) were cut, transferred to a No.
l'/2 coverslip, and teased apart with two pairs of fine for-
ceps in a drop of saline solution. The fragmented retina
preparation was covered with a second No. 1 Vi coverslip
of smaller size, blotted gently along its edges, and sealed
with a mixture of molten paraffin and Vaseline, as de-
scribed earlier (Harosi and MacNichol. 1974a). The saline
we used here was a modified marine teleost Ringer so-
lution (Forster and Taggart, 1950), containing 10 mA/
HEPES buffer at pH 7.3.
Spectrophotometer
The spectral measurements were carried out with the
help of the dichroic microspectrophotometer (DMSP) de-
scribed previously (Harosi and MacNichol, 1974b; Harosi,
1982, 1987). The DMSP is a computer-controlled, wave-
length-scanning, single-beam photometer that records
transmitted light fluxes through microscopic samples. The
measuring beam is commonly adjusted to about 1 X 3
nm in the plane of the specimen, and its spectral purity
(monochromaticity) to about 5 nm. This beam is focused
by a quartz field lens onto the back aperture of the con-
denser through a Glen-Thompson UV polarizer and a
CaF: photo-elastic modulator. The condenser we routinely
used was a 32/0.4 Ultrafluar (Zeiss), whereas the objective
was a 100/1.3 UV-F100 (Nikon) microscope objective,
both of the glycerine immersion type. Due to limitations
imposed by the latter objective, light detection was possible
only at wavelengths greater than about 330 nm.
Spectral recording and analysis
Average and modulated light fluxes were detected with
a cooled photomultiplier tube (Hamamatsu. Type R375),
and photocurrents were recorded into two sets of 75 se-
quential memory locations as the wavelength was scanned
rapidly (500 nm/s) from the short wavelength to the long
wavelength end of the spectrum (275-645 nm). Corre-
sponding signals were summed, and the memory locations
thus contained numbers signifying transmitted flux am-
plitudes averaged over 5-nm-wide segments of the spec-
trum. A typical measurement included 16 background
scans from a cell-free area in the preparation (reference
measurement). 8-16 prebleach scans (sample measure-
VISUAL PIGMENTS IN BLACK SEA BASS
137
ment). a 2-min exposure to actinic light provided by the
measuring beam (the wavelength of which was preset to
the vicinity of the expected Xmax) if bleaching was desired,
and a 16-scan postbleach recording of the sample. The
dedicated digital computer of the DMSP subsequently
calculated (from the average and modulated transmitted
fluxes) the average absolute absorbance (A), the bleaching
difference absorbance (BD). and the linear dichroism (LD)
spectra. Absorbance (optical density) is denned as logdT '.
where T is trr remittance. Linear dichroism is proportional
to sample polarization, denned as p = (Tn - - T J/(T
+ T\). The LD ordinate is calibrated so that a perfect
analyzer would yield +1, if crossed, and —1, if parallel to
the plane of the polarizer. For details of the measurement
technique, the selection of spectra, and data analysis, see
Harosi(1975a, 1987).
Visual pigment characterization
Our spectroscopic description of a pigment is based on
A. BD, and LD determinations from optically isolated
single or overlapping multiple photoreceptor cells. Em-
pirical evidence suggests that, in general, the three types
of measurement yield three Xmax values that will "bracket"
the "true" Xmax of the visual pigment. Occasionally. Xmax
may be slightly blue-shifted (due to photoproduct ab-
sorption and excessive short-wave scattering); BDmax is
usually red-shifted (because shortwave absorbances. which
are subtracted, tend to be exaggerated): whereas LDmax
should theoretically be close to Xmax (provided there is no
instrumental delay between the "ac" and "dc" detection
channels). Rhodopsins (based on the aldehyde of vitamin
A! . or retinal) and porphyropsins (based on the aldehyde
of vitamin A2. or 3-dehydroretinal) have several distin-
guishing properties /;; situ: ( 1 ) «-band half-bandwidth
(HBW) value (i.e., rhodopsins being narrower than por-
phyropsins); (2) /j-band absorbance (i.e.. rhodopsins hav-
ing relatively lower /3-band absorbance than porphyrop-
sins): (3) transverse specific density (i.e., rhodopsins have
higher molar extinction than porphyropsins); (4) dichroic
ratio (i.e.. rhodopsin bearing cells show greater optical
anisotropy than those with porphyropsin); and (5) absolute
Xmax value, which is informative only beyond 570 nm
(i.e.. no purely retinal-based pigment has ever been found
with Xmax greater than about 570 nm).
Following Fourier smoothing of the "raw" spectra,
software algorithms can determine the peak absorbance
and half-bandwidth values. The a-band of a typical rho-
dopsin absorbance spectrum has a HBW of 4000-4100
cm" ' . whereas that of a typical porphyropsin is about 4800
cm'1. Moreover, in both classes of visual pigment, the
HBW is a function of wavelength, such that with increas-
ing Xmax. the HBW progressively narrows, and with de-
creasing Xmax. it progressively broadens. We made use of
these properties in our characterization of the sea bass
visual pigments.
Quantal absorption of pigments vs. environmental
illumination
In search of the correspondence between visual pig-
ments and the photic environment, we analyzed models
that are consistent with the premise that photoreceptors
are quantum detectors with a response primarily depen-
dent upon the total number of quanta absorbed per unit
time by their visual pigment (rate of quantum catch),
bleaching and regeneration notwithstanding. We further
assumed that downward irradiance of solar origin is the
primary determinant for the sea bass visual system. We
made use of data available in the literature on solar ir-
radiance and on the optical properties of natural bodies
of water, and we used our own spectroscopic determi-
nations on the photoreceptors.
We performed the following calculations. ( 1 ) Using the
standard solar irradiance data of Moon (1940; Table III),
by a procedure similar to that adopted by Dartnall ( 1 975).
we generated quanta! irradiance values expected at sea
level in quanta/s X mm2 X nm. at 1-nm intervals (by
linear interpolation). (2) The classification on optical water
types and transmittance data of Jerlov (1968: Table XX)
permitted the transformation of downward irradiance
values to any depth. Following the suggestion of Dartnall
(1975), his five oceanic water types were designated JI.
JIA, JIB. JII, and JIII, and the five coastal water types as
Jl, J3, J5. J7. and J9. Downward irradiances were cal-
culated for all 10 water types for depths of 10, 20, 50,
100, 150, and 200 m for 1-nm intervals. (3) The visual
pigment absorbance spectra obtained at 5-nm intervals
were again interpolated to 1 nm in the available range of
350-650 nm. (4) Rate of quantal absorption (quantum
catch) by each visual pigment was determined for the three
receptor types at six depths in ten water types. The total
absorbed quantum flux density rate Q, was. in each case,
obtained by summing the products of the appropriate
quantal irradiance and receptor absorptance at each
nanometer of wavelength. Absorptance had the usual def-
inition: A(X) = 1 - lO'1**', where D(X) = Dmax Are, (X).
The peak absorbance. Dmax, was obtained as a product
S^ (Table I) and the axial pathlength through the outer
segment of the receptor type containing the pigment. The
average lengths observed in our video records for single
cone. rod. and double cone outer segments were 9, 20,
and 23 nm. respectively. Arel (X) was derived from the
normalized absorbance spectra (depicted in Fig. 5).
In an attempt to find criteria by which the correspon-
dence between visual pigment absorptance and environ-
mental light could be judged, we calculated the wave-
lengths at which 25%, 50%, and 75% of the total quantum
catch occurs in each receptor type in a given photic en-
vironment. With a symbolic designation of Xqc50 for the
50% value, this is analogous to the XP50 introduced by
Munz and McFarland (1973). Note, however, the differ-
138
Spectral data for black sea bass cones
K. V. SINGARAJAH AND F. I. HAROSI
Table I
No. of
HBW Sj.
Cone type
determ.
\na« [nm]
[cm
'] AmM[OD] R d^m]
[OD/cm]
Single
6
0.03167 ±0.0065 1.64 ±0.1 9 2.9 + 0.3
136 ±22
A
6
463.2 ± 2.2
4638 ±
296
LD
6
460.3 ± 2.3
4670 +
625
BD
3
469 ± 3.5
3737 ±
673
A + LD
12
461.7 ± 2.7
4654 +
467
A + LD + BD
15
463.2 ± 4
4470 ±
616
Best Estimate
463 ± 4
4500 ±
600
Double
8
0.04774 ±0.0160 1.52 + 0.14 3.0 ± 0.4
161 + 39
A
8
526.8 ± 5.5
3928 +
197
LD
8
527 ± 4.7
3728 +
237
BD
5
529.7 ± 4.5
3634 ±
159
A + LD
16
526.9 ± 5
3828 +
235
A + LD + BD
21
527.5 ± 4.9
3782 ±
231
Best Estimate
527 ± 5
3800 ±
200
Abbreviations used: HBW, halt-bandwidth; Amaj, peak absorbance; OD. optical density: R, dichroic ratio; d, mean diameter of outer segment; Sj.,
specific density (transversely polarized): A, «-band of absorbance spectrum; LD, a-band of linear dichroism spectrum; BD, a-band of bleaching
difference absorbance spectrum.
ence between the two: while XP50 is the wavelength at
which 50% of all the quanta occur in the spectrum of
400-700 nm in a given photic setting, Xqc5(, is a measure
of the absorptive interaction between light and pigment
throughout their available spectral range.
We also calculated total quantum catch ratios between
the receptor types in the retina of this fish. Although each
Qt value is critically dependent on the axial pigment den-
sity assumed in the calculation, the Q, ratios indicate the
relative "weight" of the receptor types in the retina. As
the spectral distribution of light changes with water type
and with depth, the quantum catch by the receptor types
vary, and the Q, ratios may go "out of balance." Although
we do not know the range of appropriate quantum catch
ratios, we found them to be indicative of the "spectral
match" that exists between photoreceptors and environ-
mental light. The second criterion by which to assess the
appropriateness of a visual pigment to a given environ-
ment is the difference between Xqc50 and the Xmax of the
pigment in question. Again, we do not know how large
this difference should be before it becomes unacceptable.
Based on the experience gained in this analysis, we ten-
tatively set the limit of acceptability at 30 nm for the
difference between Xqc50 and Xmax of a pigment, and 100%
for the change in Q, ratio of two receptor types.
We made no attempt to account for the retinal distri-
bution of the different pigments, their relative proportions,
or the light collecting efficiency of the various cell types.
Nor did we consider the light collection efficiency of the
eye as a whole, or the losses of light that occur at the
ocular media and their interfaces. The modifying effects
of these factors we envisage investigating in the future as
sufficiently accurate and detailed information becomes
available.
Results
Although we measured well over one hundred photo-
receptors, permanent records were saved from 1 7 single
cones and 66 outer segments belonging to 39 double cones.
Additionally, we recorded from 12 groups of multiple
rods. In our fragmented retina preparations, the most fre-
quently occurring cells were rods. While double cones
could also be located with relative ease, single cones were
less numerous. Photoreceptor morphology is illustrated
in Figure 1. With these examples, we wish to make the
point that double cones were variable in size and shape:
the two members were practically identical in some, but
quite different in others. For this reason, we simply refer
to them as "double" and refrain from the use of the term
"twin," even though the visual pigments in the two mem-
bers were spectroscopically indistinguishable, as will be
shown below.
Single cones
All single cones encountered had one type of shortwave,
or blue-absorbing, visual pigment. The outer segments
were transversely dichroic, and their pigment content was
bleachable. Representative spectra are depicted in Figure
2. Spectral data are summarized in the upper part of Table
I; note the several subgroup averages calculated for Xmax
and HBW. The "best estimate" for each is based on the
over-all average.
VISUAL PIGMENTS IN BLACK. SEA BASS
139
Figure 1. Black sea bass photoreceptors viewed in infrared illumination in the recording microspectro-
photometer. The images were captured on video tape and subsequently photographed from a video monitor
display. A. Double cone in lateral view, just below a rod outer segment. B. Double cone with unequal outer
segments. C. Double cone in a rare orientation, with overlapping outer segments. D. Single cone, proximal
to retinal fragments. All four panels have equal magnification and the bar length represents 10 Mm.
Double cones
Every outer segment belonging to double cones was
transversely dichroic due to the presence of a bleachable
pigment. Representative spectra are shown in Figures 3
and 4, and spectral data are presented in the lower part
of Table I. The A and BD spectra of Figure 3 were derived
from one member of double cones. The spectra of Figure
4 were obtained from overlapping outer segments of dou-
ble cones (see panel C of Fig. 1 ). Note the increased spec-
tral absorbance in Figure 4A as compared to that in Figure
3A. The idea we illustrate here is that, when the two outer
segments overlap laterally, the transversely scanned ab-
sorbance nearly doubles, as it should if the two members
are equivalent. However, while the A and BD spectra in-
crease in proportion when measured from two, instead
of one member, the Xmax and the HBW remain virtually
unchanged. This can happen only if the same pigment is
contained in both members. There was no evidence for
the presence of a second pigment in any of the double
cones.
Rods
As is common in teleost retinas, the rods of the black
sea bass are numerous; the outer segments are of variable
length and slender, with a diameter of 1 ^m or less. Re-
cordings from multiple rods yielded A, BD, and LD spec-
tra indicative of a "typical" rhodopsin. Trace B in Figure
5 was derived from such absorbance spectra. The HBW
of the tt-band of the rod absorbance spectra were within
experimental error of the value obtainable from other
rhodopsins (such as amphibian or monkey), and this pig-
ment should, therefore, also be based on vitamin A,.
Traces A and C in Figure 5 depict the normalized absor-
bance spectra of the cone pigments. The two cone pigment
spectra flank the rod pigment spectrum on the longwave
and shortwave sides by nearly the same distance on the
wavelength scale.
Dichroic ratio and transverse specific density
The algebraic relationships necessary to determine cel-
lular dichroic ratios from the A and LD spectral mea-
140
K. V. SINGARAJAH AND F. I. HAROSI
m
u
c
o
.£>
_O
H
— I —
tee
Wavelength (nm)
0.02-
-Q
b
0 -0.01-
— i 1 1 —
450 500
1 1 1 —
550 600
Wavelength (nm)
0.02-
y
350 400 450 500 550 600 650
Wavelength (nm)
Figure 2. Absorbance, bleaching difference absorbance, and linear
dichroism spectra of the visual pigment in single cones of the black sea
bass. A. Average absorbance spectrum based on four single-cell recordings
( + ). The solid curve is the result of Fourier smoothing. Peak absorbance
and half-bandwidth are 464 nm and 4430 cm"', respectively. B. Bleaching
difference spectrum from one cell. Data values (O) were derived from
prebleach and postbleach measurements consisting of the sum of 16
scans, each. The dashed curve is based on Fourier-smoothed data. The
positive band peaks at 471 nm, with HBW = 3140 cm"1; the negative
band peaks at 387 nm, with HBW = 3570 cm '. C. Average linear
dichroism from 3 cells (A). The dotted line, again, is the result of Fourier-
surements have been published previously (Harosi, 1987).
The results for the black sea bass cones are listed in Table
I. The last column in Table I shows numerical values for
the transverse specific density, Si. For this determination,
the transversely polarized component of the peak absor-
bance, AI, is needed. The latter is derived from the average
(unpolarized) peak absorbance. A, divided by factor f,
which in turn depends on the dichroic ratio, R, defined
as R = AI/AI,. Thus, f = ( 1 + R)/2R and Ax = A/f.
Finally, S± = Ai/d, where d is the mean diameter of the
compartment (Retry and Harosi, 1990). Thus, the mean-
ing of S± is peak absorbance for transversely polarized
light per unit thickness (measured either in micrometers
or centimeters).
Discussion
In our teased preparations obtained from various re-
gions of the retina of the black sea bass, we found rods,
double cones, and single cones. The outer segment in each
of these cells contained a visual pigment characterized by
a typical absorption spectrum, dichroism, and light-sen-
sitive spectral changes (i.e.. bleaching). On the basis of
half-bandwidth determinations, we identified the chro-
mophore of these pigments as retinal. Additional evidence
comes from the low /3-band absorbances which we com-
monly observed. The results on specific density (Si) also
support this, because the cones yielded higher values of
this parameter than those obtainable from cells using pig-
ments with 3-dehydroretinal as chromophore, although
not quite as high as what has been reported in cases of
amphibians and monkeys. This discrepancy may be re-
lated to the results on dichroic ratio, which were also below
expectation (see below).
The presence of only two cone pigments would make
the black sea bass dichromatic in the traditional sense,
although we have no evidence which could preclude the
rods from chromatic discrimination tasks. Nevertheless,
"color vision" can be supported by only two cone mech-
anisms, as we know from other studies on animals, as
well as on humans. The existence of many vertebrates
with trichromatic and even tetrachromatic cone mecha-
nisms raises the question as to why this species has evolved
only two. The simplest answer is, perhaps, that there was
no selective pressure to have more. Given the relatively
narrow spread of wavelengths at greater depths, there is
probably no advantage in having more cone types, even
though the eye's spectral resolution could be improved
by adding more closely spaced "color channels." While
vertebrate eyes would make very poor spectrographs, they
nonetheless serve the bearer well. To evaluate just how
well an organism is served by its eyes, we would need to
know not only the lighting conditions and reflectance
properties of all objects, but also the visual tasks that need
to be solved in capturing food, avoiding predators, finding
VISUAL PIGMENTS IN BLACK. SEA BASS
141
mates, and continuing successful reproduction (Dartnall,
1975; Levine and MacNichol, 1979). Clearly, more
knowledge is required before precise answers can be found.
Similar visual pigments have been previously reported
for four other species of fish. Also using microspectro-
photometry, Loew and Lythgoe ( 1 978) investigated several
species offish from various "environmental groups." In
the "moderately deep coastal group" they reported finding
two cone pigments with Xmax of 460 and 530 nm, respec-
tively, and a 502-nm rod pigment in two species of gurnard
( Trigla lucerna and Eutrigla giirnardus). Two other species
of marine fish with similar pigments were found by Levine
and MacNichol (1979). These were the sea robin (Prio-
notus carolinns) and the scup (Stenotomus versicolor).
Photic habitat of the black sea bass
Although information on habitat is rather scanty, this
fish inhabits waters within a depth range of a few meters
from the surface, to 165 m. Being demersal, they are
caught in large numbers in waters of 50-150 m depth.
This species is mainly a bottom feeder, and prefers to be
among rocks and reefs. Males have been observed to de-
velop a bright blue color prior to spawning. Spawning
involves buoyant eggs in depths ranging between 18-45
m: the larvae tend to move to inshore waters over rocky
bottoms (Bigelow and Schroeder, 1953; Perlmutter, 1961;
Gordon, 1977).
Relevant spectral data on habitat
There is a general dearth of information, particularly
field measurements, on the photic environment of the
black sea bass. The specimens we used were caught in
Woods Hole Harbor, where the color of the water is green.
This agrees with Clarke and Denton ( 1 962), who reported
that the maximum transparency of coastal waters can be
generally found in the range of 500-600 nm. We think
that type 5 of the coastal series of Jerlov (J5) would be
appropriate for the optical characterization of this habitat.
Because the black sea bass is a widely distributed species,
ranging from southern Massachusetts to Florida and from
bays and sounds to Georges Bank, it will encounter off-
shore oceanic waters as well. Based on Jerlov's (1968)
regional classification of optical water types, the western
North Atlantic is described by type IB of the oceanic series
(JIB). Thus, it appears reasonable to assume, as initial
conditions, that this fish's visual system needs to cope
with photic habitats expected of optical waters from JIB
to J5. But these a priori assumptions are unnecessary, for
similar conclusions can be drawn from the analysis dis-
cussed below.
Correlation between downward irradiance and receptor
pigments
We calculated the rate of quantum flux density ab-
sorption ("quantum catch") by the sea bass visual pig-
.07 -
S .0"
I .03
L
o
I/I .02
-D
ff .11
450 5BB 550
Wavelength (nm)
_D B.03
O
„» »
Wavelength (nm)
Figure 3. Absorbance and bleaching difference absorbance spectra
derived from one of the outer segments of a double cone in the black
sea bass. A. Average absorbance obtained in 16 scans (+). The solid
curve represents the result of Fourier smoothing; Xraa;l = 524 nm. HBW
= 4060 cm"1. B. Average of three bleaching difference spectra, each of
which is based on 16 scans (O); the dashed curve is derived from Fourier
smoothing. The positive band peaks at 527 nm. with HBW = 3590
cm"1, whereas the negative band peaks at 395 nm, with HBW = 4300
ments in the downward solar irradiance (from Moon's
data) transformed by the optical water types of Jerlov (see
Methods for details). We also determined quantum catch
(Qt) ratios between the receptor types and the wavelength
of 50% quantum catch (Xqc50) for each pigment at each
depth. Table II depicts some of the results. Overall, the
data show good agreement between Xqc50 and Xmax of the
blue pigment in ocean waters and the same for the green
pigment in coastal waters. Although the absolute values
of Q, and Xqc50 vary for the three receptor types in the
depth range of 10-200 m and across the oceanic types JI-
JIII. the Q, ratios show no drastic variations. Nor do the
Xqc50 and Xmax differences. This is not true for the coastal
water types. Even at 10 m depth, these parameters undergo
significant changes from Jl to the other types, so that in
142
K. V. SINGARAJAH AND F. I. HAROSI
S .04
I .03
L.
S .02
JD
* .0,
-.01
Wavelength (nm)
_c
u 0.02-
— 1 1 1 1 1 1 1 1 1 1 I r—
3SB 400 450 500 550 600 650
Wavelength (nm)
Figure 4. Absorbance and linear dichroism spectra obtained simul-
taneously from both outer segments (stacked one above the other) of a
double cone in the black sea bass. A. Average absorbance. based on 16
scans ( + ); the solid curve is the product of Fourier smoothing: Xmax
= 521 nm, HBW = 4010 cm"'. B. Linear dichroism from the corre-
sponding structure, based on 16 scans (A); the dotted curve is the outcome
of Fourier filtering. The Xmax and the HBW are 525 nm and 3860 cm"1,
respectively.
J7, for example, G/B = 10.2, and in J9 it is 26.8. With
increasing depth in these types of water, the sea bass visual
pigments are clearly out of tune. For the J5 water type,
the parameters are probably within acceptable range to
depths of 20 m (see Table II). At greater depths, however,
the fitness of the pigments become questionable. For in-
stance, in J5 at 50 m depth the Xqc50 and Xmax differences
are 55.6 nm, 27.5 nm, and 2.3 nm for the B, R, and G
receptor types, respectively; the absolute Q, values are
down six orders of magnitude with respect to those at
10 m, and the R/B, R/G, and G/B ratios yield 5.3, 0.7,
and 7.5, respectively. Although we lack firm criteria by
which to interpret these numbers, they probably indicate
an intolerable mismatch of the pigments to this photic
habitat.
On the magnitude of the dichroic ratio in cones
Ever since the discovery of linear dichroism in laterally
viewed rods by Schmidt ( 1938), the phenomenon has been
interpreted in terms of a structural anisotropy in the outer
segments of vertebrate photoreceptors. A quantitative
measure of this property is the cellular dichroic ratio, R,
as was defined in an earlier section. The magnitude of R
is an expression of structural "order"; i.e., the larger the
R, the more ordered is the disposition of the visual pig-
ment in the cell. Aside the complexities of interpretation,
rhodopsin-containing rods yield larger R values than por-
phyropsin-bearing rods (Harosi, 1975b). Furthermore,
cone dichroic ratios are always smaller than those obtain-
able from rods, regardless of pigment class. In published
accounts, for example, goldfish (Harosi and MacNichol,
1974a), Japanese dace (Harosi and Hashimoto, 1983), and
carp (Hawryshyn and Harosi, 1991) yielded mostly values
with R > 2. In comparison, the average values we obtained
for the black sea bass were Rouble = 1.52 ± 0.14 (n = 8)
and Rs,ng]e = 1.64 ± 0.19 (n = 6) (see Table I). Although
in one instance R = 1.9 was found in a single cone, we
conclude that the values are rather smaller than they ought
to be. Obviating the trivial interpretation of these results
as instrumental artifact, there is the possibility that our
experimental specimens were subnormal in their photo-
receptors. The notion we entertain here is that these fish,
although appearing quite healthy, nonetheless have suf-
fered subtle structural damage in their retinal receptors
1.10
1.00
0.90-
0.80-
0.70-
0.60-
0.50-
350 400 450 500 550 600 650
Wavelength (nm)
Figure 5. Comparison of relative absorbance spectra determined from
the three visual pigments present in the black sea bass. Each curve was
obtained from an experimental absorbance spectrum by dividing the
data set by its peak value. The dotted curve (A) is based on Figure 1A
The solid curve ( B) was derived from two sets of multiple rod absorbance
measurements, each consisting of 16 scans. The dashed curve (C) was
replotted from the spectrum of Figure 2A. The Xmal and the HBW of
these spectra are: for single cone, 464 nm, 4430 cm~'; for rod, 498 nm,
4260 cm"1; for double cone. 524 nm, 4060 cm"1, respectively. Note that
the two outer segments of double cones contained visual pigments that
were spectroscopically indistinguishable from one another.
VISUAL PIGMENTS IN BLACK SEA BASS
143
table II
Ahsorhctl quantum flux. density rules hy black sea buss visual pigments, Q:. in 10'~ quanla/s x mm~ {integrated in I-nm steps from 350-650 nm)
as a Junction ol depth
Single cone
Rod
Double cone
Depth
[m]
163 nm
XqcSO [nm]
Q,
= 498 nm
\cso [nm]
G, Xma, — 527 nm
Qt \,c5o [nm]
B, Ama» - '
Q,
R/B
R/G
G/B
For oceanic
water type JIB:
10
40.08
472.0
111.49
491.0
122.32
509.2
2.782
.911
3.052
20
27.15
470.0
71.22
486.8
73.18
501.3
2.624
.973
2.696
50
8.762
468.4
20.49
478.8
19.04
487.9
2.338
1.076
2.172
100
1.426
466.1
3.022
472.9
2.630
477.7
2.119
1.149
1.844
150
.2435
465.1
.4901
470.1
.4148
473.5
2.013
1.182
1.704
200
.0427
464.5
.0835
468.6
.0696
471.2
1.953
1.199
1.629
For coastal water type J5:
10
20
1.106
.0319
499.1
507.9
4.429
.1462
516.7
521.5
6.248
.2088
529.1
529.8
4.005
4.580
.709
.700
5.651
6.541
Notes:
(1) B and G are blue- and green-absorbing visual pigments, while R stands for rhodopsin of the rod.
(2) Absorbed quantum flux density was obtained from absolute absorptance of a pigment by the relation A (X) = 1 — lO"0**', where D (X) is optical
density at a given wavelength. D (X) was derived from the relative absorbance values (spectra shown in Fig. 5) multiplied by the peak axial absorbances
of the pigments in situ. Actual peak absorbances used for the single cone, rod, and double cone pigments were 0.12, 0.32. and 0.37, respectively.
(3) JIB and J5 correspond, respectively, to the optical water types IB and 5 of Jerlov (1968).
(4) Xqc50 is the wavelength at which 50% of the "total quantum catch" is attained by a visual pigment in a given photic environment.
due to the artificial illumination (and possibly from nu-
tritional deficiency) while in captivity. In view of a number
of recent studies of light damage conducted primarily on
mammals, the idea has rational basis and appears worthy
of further investigation.
Summary
By determining the visual pigments in the retina of the
black sea bass and analyzing the expected quantum catch
by the photoreceptor types, we obtained indications for
the preferred photic habitat. We used two criteria to assess
the fitness of a set of visual pigments to ambient light: ( 1 )
ratio of quantum catches by the receptor types; and (2)
the deviation of the wavelength at 50% quantum catch
from the Xmax of each pigment in a given photic environ-
ment. Indications were that the black sea bass has a bal-
anced set of three pigments to match the downward ir-
radiance spectrum of clear ocean water to depths of 200
m. However, for coastal waters the fit is limited to the
optical types Jl, J3, and J5, and the last type to only a
depth of about 20 m. Water types J7 and J9 are expected
to be unsuitable to the black sea bass at all but the shal-
lowest of depths. This suggests that the Xmax of visual pig-
ments in a multi-pigmented system are selected on the
basis of the photic interaction between environmental light
and the pigment spectrum to produce a balanced quantum
catch in the receptor types. Whether this is a general rule.
or there are other criteria by which one Xmax is preferred
over another, are questions to be settled by future inves-
tigations.
Acknowledgments
We wish to express our appreciation to Fred Nichy and
his colleagues of the National Marine Fisheries Service,
NOAA, Woods Hole Laboratory for their unfailing sup-
port in providing us with the experimental specimens for
this work. Financial support was provided by the Federal
Government of Brazil — CAPES, Ministry of Education,
while one of us was on sabbatical leave (KVS) and grant
EY04876 of the USPHS (to FIH).
Literature Cited
Bigelow, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine.
Fishery Bulletin of the Fish and Wildlife Service, Vol. 53. U. S. Gov-
ernment Printing Office. Washington, DC. 577 pp.
Clarke, G. L. 1936. On the depth at which fish can see. Ecology 17:
452-456.
Clarke, G. L., and E. J. Denton. 1962. Light and animal life. Pp. 456-
468 in The Sea. Vol. 1 . Physical Oceanography. M. N. Hill, ed. In-
terscience. New York.
Dartnall, H. J. A. 1975. Assessing the fitness of the pigments for their
photic environments. Pp. 543-563 in I'ision in Fishes: New Ap-
proaches in Research. M. A. Ali, ed. Plenum Press, New York.
Denton, E. J., and F. J. Warren. 1957. The photosensitive pigments
in the retinae of deep-sea fish. J. Mar. Biol. Assoc. (U. A'.,1 36: 651-
662.
144
K. V. SINGARAJAH AND F. I. HAROSI
Forster, R. P., and J. V. Taggart. 1950. Use of isolated renal tubules
for the examination of metabolic processes associated with active
cellular transport. / Cell Comp. Physiol. 36: 251-270.
Gordon, B. L. 1977. The Secret Lives of Fishes. Grossarl & Dunlap
Publishers, New York. 305 pp.
Harosi, F. 1. 1975a. Microspectrophotometry: the technique and some
of its pitfalls. Pp. 43-54 in Vision in Fishes: New Approaches in
Research, M. A. All, ed. Plenum Press, New York.
Harosi, F. I. 1975b. Absorption spectra and linear dichroism of some
amphibian photoreceptors. J. Gen. Physiol. 66: 357-382.
Harosi, F. I. 1982. Recent results from single-cell microspectropho-
tometry: cone pigments in frog, fish and monkey. Color Res. Applic.
7 (No. 2. Part 2): 135-141.
Harosi, F. I. 1987. Cynomologus and rhesus monkey visual pigments:
application of Fourier transform smoothing and statistical techniques
to the determination of spectral parameters. J. Gen. Physiol. 89: 717-
743.
Harosi, F. I., and Y. Hashimoto. 1983. Ultraviolet visual pigment in
a vertebrate: a tetrachromatic cone system in the dace. Science 222:
1021-1023.
Harosi, F. I., and E. F. MacNichol, Jr. 1974a. Visual pigments of
goldfish cones: spectral properties and dichroism. J Gen. Physiol.
63: 279-304.
Harosi, F. I., and E. F. MacNichol, Jr. I974b. Dichroic microspec-
trophotometer: a computer-assisted, rapid, wavelength-scanning
photometer for measuring linear dichroism in single cells. J. Opt.
Soc. Am. 64:903-918.
Hawryshyn, C. \V., and F. I. Harosi. 1991 . Ultraviolet photoreception
in carp: microspectrophotometry and behaviorally determined action
spectra. I 'ision Res. 31: 567-576.
Jerlov, N. G., 1968. Optical Oceanography. Elsevier Publishing Co.
Amsterdam, 194 pp.
Irvine, J. S., and E. F. MacNichol, Jr. 1979. Visual pigments in teleost
fishes: effects of habitat, microhabitat. and behaviour on the visual
system evolution. Sensory Processes 3: 95-131.
Loew, E. R., and J. N. Lythgoe. 1978. The ecology of cone pigments
in teleost fishes. Vision Res 18: 715-722.
Lythgoe, J. N. 1972. The adaptations of visual pigments to the photic
environment. Pp. 566-603 in The Handbook of Sensory Physiology.
Vol. VII/1, Photochemistry of 'Vision, H. J. A. Dartnall, ed. Springer,
New York.
Lythgoe, J. N. 1984. Visual pigments and environmental light. Vision
Res. 24: 1539-1550.
McFarland, W. N., and F. W. Munz. 1975a. Part II: The photic en-
vironment of clear tropical seas during the day. I 'ision Res. 15: 1063-
1070.
McFarland, W. N., and F. W. Munz. 1975b. Part III: The evolution
of photic visual pigments in fishes. Vision Res. 15: 1071-1080.
McFarland, \V. N., and F. W. Munz. 1975c. The visible spectrum dur-
ing twilight and its implications to vision. Pp. 249-270 in Light as
an Ecological Factor, G. C. Evans, R. Bainbridge and O. Rackham,
eds. Vol. 2. Blackwell, Oxford.
Moon, P. 1940. Proposed standard solar-radiation curves for engi-
neering use. J. Franklin Inst. 230: 583-617.
Munz, F. \V. 1958. Photosensitive pigments from the retinae of certain
deep sea fishes. J. Physiol. 140: 220-225.
Munz, F. \V. 1964. The visual pigments of epipelagic and rocky shore
fishes. Vision Res 4: 441-454.
Munz, F. VV., and VV. N. McFarland. 1973. The significance of spectral
position in the rhodopsins of tropical marine fishes. Vision Res. 13:
1829-1874.
Perlmutter, A. 1961. Guide to Marine Fishes. New York University
Press. 431 pp.
Petry, H. M., and F. I. Harosi. 1990. Visual pigments of the tree shrew
( Tupaia helungcri ) and greater galago (Galago crassicaiidatus): a mi-
crospectrophotometric investigation. Vision Res. 30: 839-851.
Schmidt, W. J. 1938. Polarisationsoptische Analyse eines EiweiC-Li-
poid-Systems, erlautert am AuBenglied der Sehzellen. Kolloid-Z. 85:
137-148.
Schultze, M. 1866. Zur Anatomic und Physiologic der Retina. Archiv
Mikrosk. Anal. 2: 175-286.
Schultze, M. 1867. Ueber Stabchen und Zapfen der Retina. Archiv
Mikrosk. Anal. 3:215-247.
Reference: Biol. Bull. 182: 145-154. (February, 1992)
Quantitative Analysis of the Structure and Function of
the Marsupial Gills of the Freshwater Mussel
Anodonta cataract a1
RICHARD A. TANKERSLEY AND RONALD V. DIMOCK, JR.
Department of Biology, Wake Forest University, Winston-Salem, North Carolina 27109
Abstract. Gravid females of Anodonta cataracta incu-
bate shelled larvae (glochidia) in the water tubes of their
outer demibranchs which, in turn, undergo extensive
morphological changes in becoming marsupia. In this
study, the brooding gills of A. cataracta were compared
to the non-marsupial demibranchs of females and the gills
of males. Scanning electron microscopy and video en-
hanced light microscopy were used, and computer-gen-
erated 3D-reconstructions of gill tissue were also prepared
from light micrographs of serial histological sections.
Marsupial gills possess a tripartite system of water tubes
that are not present in non-marsupial gills and include
two secondary water channels and one primary water tube
(brood chamber) containing glochidia. The lateral di-
mension (width) of water tubes of the marsupial gills in-
creases nearly 30-fold during brooding, but the anterior-
posterior length of the tubes is unaffected. No apparent
changes in the morphology of the non-marsupial inner
demibranchs were observed. Glochidia are effectively iso-
lated from the surrounding water by secondary septa, po-
sitioned between the primary and secondary water tubes.
Secondary septa are present during brooding and im-
mediately after larval release, but are not in evidence
Received 15 July 1 99 1 ; accepted 1 1 October 1991.
1 Contribution #297 from the Tallahassee. Sopchoppy and Gulf Coast
Marine Biological Association.
Abbreviations: ANOVA: analysis of variance: AV, arterial vessel; BC.
brood chamber, CE. ciliated water tube epithelial cell; F, gill filament;
FT, foot: G, glochidmm; ID, inner demibranch; IFC, interfilament water
canal; ILS. interlamellar septum; LT, lamellar tissue: MP, melting point;
N. nerves; O. ostium; OD, outer demibranch; PWT, primary water tube;
SEM, scanning electron microscopy; SS, secondary septa; SWT, sec-
ondary water tube; 3D, three dimensional; PCI & PC2. principal com-
ponents I & 2; PCA. principal components analysis; VM. visceral mass.
among females during non-reproductive periods. Quan-
tification by 3D reconstruction revealed that, although
secondary water tubes are smaller than the primary water
tubes of non-marsupial gills and non-gravid marsupial
gills, collectively they provide about the same cross-sec-
tional area as the primary water tubes that are lost to
water transport by occlusion with glochidia. However,
considering the fluid dynamics of the ciliary gill pump,
net water transport through the lumina of marsupial gills
is reduced to only about 16% of that in non-gravid mar-
supial demibranchs.
Introduction
Unlike their marine counterparts, most freshwater bi-
valve mollusks, including the Sphaeridae and Unionidae,
lack a planktonic larva and bypass the trochophore and
veliger stages; rather, they incubate their embryos, larvae,
or both in their gills. Moreover, the life cycles of the
Unionidae are atypical among bivalves in including both
a free-living adult and a short-lived obligatory ecto-par-
asitic larval (glochidial) phase (Coker et ai, 1921; Kat,
1984). Following fertilization in the suprabranchial cavity,
embryos develop in the water tubes of the female's gills.
During reproduction, both outer (lateral) demibranchs of
Anodonta cataracta (recently reassigned to the genus Py-
ganodon by Hoeh, 1990) serve entirely as a pair of mar-
supial chambers and undergo pronounced morphological
and architectural changes to accommodate nearly a mil-
lion developing larvae (Fig. 1 ). Anodonta cataracta is a
dioecious long-term (bradytictic) brooder that spawns in
the late summer, broods throughout the fall and winter,
and releases mature glochidia in the early spring (Tank-
ersley, unpub. data).
General descriptions of the gill structure and anatomy
of several unionid species, including .4. cataracta, and the
145
146
R. A. TANKERSLEY AND R. V. DIMOCK. JR.
Figure 1 . Scanning electron micrograph of a frontal cross section of
the marsupial gill of A twdonta cataracta showing the position of the
glochidia larvae (G) in the brood chambers and the location of the sec-
ondary water tubes (SWT) and interlamellar septa (1LS). Additional ab-
breviations: F, gill filament.
role of demibranchs as sites of larval storage during re-
production have been reported previously (Peck, 1877;
Lefevre and Curtis, 1910; Ortmann, 191 1; Richard et al,
1991). The non-marsupial gills (outer and inner demi-
branchs of males and inner demibranchs of females) of
A. cataracta possess continuous interlamellar septa that
run dorso-ventrally at right angles to the gill surface and
form evenly spaced, uninterrupted water tubes (Fig. 2)
(Ridewood, 1903; Heard and Guckert, 1971). The primary
water tubes of the marsupial lamellae are more numerous
than those in non-marsupial gills and during gravid pe-
riods are divided into three separate compartments: a
central brood chamber serving as an ovisac, and two tem-
porary secondary water tubes located on both the lateral
and medial ends of the brood chamber parallel to the
surface of the demibranchs (Fig. 2). These secondary water
tubes are formed from extensions of the interlamellar septa
prior to larval incubation, and may be associated with
the long brooding period of this species; in particular,
they are thought to be responsible for maintaining water
transport across the gill surface for respiration, filtration,
and aeration of developing larvae (Ortmann, 1911; Heard,
1975; Richard et al.. 1991).
Investigations of the flow dynamics associated with cil-
iary suspension feeding in bivalves (see Silvester and
Sleigh, 1984;J0rgensentVa/.. 1988; Silvester, 1988) have
prompted several studies of the functional anatomy and
ultrastructure of the bivalve gill (Moore, 1971; Owen,
1974; Way et al.. 1989). Although the reproductive cycles
and glochidial morphology of a variety of unionaceans
have been examined (see references in Kat, 1984, and
Gordon and Smith, 1990), few studies have documented
the changes in gill morphology associated with brooding
(Ortmann, 1911; Bloomer, 1934; Heard, 1975; Richard
et al.. 1991) or examined the functional role of the sec-
ondary water tubes as structures necessary for sustaining
water transport through the lateral demibranchs. Most
contemporary examinations of unionid gill structure and
function have focused upon the role of gills as sites for
ion transport (Kays et al.. 1990), and as storage areas for
extracellular calcium phosphate concretions (Silverman
eta/.. 1983, 1989) used during reproduction for embryonic
shell development (Silverman et al.. 1985, 1987).
The objective of the present study was to use both scan-
ning electron microscopy and video enhanced light mi-
croscopy to quantify the seasonal changes in the mor-
phology of the marsupial gills of A. cataracta females and
to compare these changes with those in the non-marsupial
inner demibranchs, and with the inner and outer demi-
branchs of males. The results indicate that, although the
marsupial gills swell to more than thirty times their non-
brooding thickness when the primary water tubes are
modified as brood chambers and are subsequently ob-
structed by incubating larvae, the construction of second-
ary water tubes partially compensates for the loss of pas-
sageways available for water transport. In addition, these
data are used to make theoretical predictions and estimates
of the influence of larval incubation on the fluid dynamics
of the gills and on their conventional roles as feeding and
respiratory organs.
Materials and Methods
Collection and maintenance of animals
Adult Anodonta cataracta were collected from Spea's
Pond, Yadkin County, North Carolina, and were main-
tained at ambient collection temperatures in glass aquaria
containing artificial pond water (Dietz and Alvarado,
1970). All mussels were sexually mature (average shell
length = 12.8 cm; range: 1 1.2-14.0 cm), were kept on a
12L:12D photoperiod for up to 10 days prior to use, and
their collections were scheduled to coincide with pre-
gravid (early July), gravid (October and December), and
post-glochidial release (late February) periods. The sex
ratio of mussels in the pond was nearly 1:1, and 100% of
the females collected during brooding periods possessed
gravid marsupia.
Preparation of gills for light microscopy and
computerized analysis of serial sections
Lateral and medial gill tissues (approximately 4 cnr)
for histological examination by video enhanced light mi-
croscopy were excised from the central part of their re-
spective demibranchs and fixed in Tissue-Fixx (Lerner
Laboratories) for 72 h. Specimens were decalcified in Cal-
Ex (Fisher Diagnostics) for 24 h, dissolving larval shell
GILL MORPHOLOGY OF ANODONTA CATAR.4CTA
147
Marsupial Gill
(Gravid)
Non-Marsupial Gill
Marsupial Gill
(Non-Gravid)
Figure 2. Schematic illustration of a cross section through a gravid female Anodonla calaracta showing
the position of the lateral and medial demibranchs, and the arrangement of the lamellar tissue (frontal
sections) of non-marsupial and marsupial demibranchs during gravid and non-gravid periods. Abbreviations:
AV. arterial vessel; BC, brood chamber. F. gill filament; FT. foot: G, glochidium larvae; ID, inner demibranch;
ILS, interlamellar septum; LT, lamellar tissue; OD, outer demibranch: PWT; primary water tube; SWT.
secondary water tube; VM, visceral mass.
and extracellular calcium concretions that might have in-
terfered with sectioning (Silverman et a/.. 1985; Richard
et al, 1991), and were then dehydrated in ethanol and
embedded in paraffin (Paraplast; MP 56°C) by vacuum
infiltration (Lipshaw Manufacturing Co.). Serial frontal
sections (7-8 ^m thick) were mounted on glass slides and
stained with hematoxylin and eosin according to the pro-
cedures outlined in Humason (1979).
Morphometric measurements (±2.0 ^m for linear and
4.0 MITT for area measurements) included the area (frontal
cross section), length (maximum anterior-posterior axis
distance), and width (maximum left-right axis distance)
of the primary water tubes (brood chambers in gravid and
post-release marsupial gills) and secondary water tubes
(gravid and post-release marsupial gills only); the gill
thickness (maximum distance between the filaments of
opposing lamellae); the thickness of lamellar tissue (max-
imum distance from the base of filaments to the lumen
of the primary or secondary water tubes); and the number
of filaments per interlamellar septum (including filaments
present on both ascending and descending lamellae).
These measurements were made with an Image- 1 Video
Image Analyzer (Universal Imaging Corp.) and a Ha-
mamatsu C2400 video camera and Javelin color video
camera attached to a Zeiss Axiophot microscope and a
Nikon SMZ-2T dissection microscope, respectively. Three
sets of measurements on every fourth section, for a total
of 1 2 sets per specimen ( 5-6 specimens/sex/collection pe-
148
R. A. TANKERSLEY AND R. V. DIMOCK. JR.
riod), were analyzed to account for any within-individual
variation.
We performed a principal components analysis (PCA)
(SYSTAT Statistical Software; Wilkinson, 1990) on the
original (log transformed) variables to derive a smaller set
of uncorrelated \ uriables based on linear combinations
of the original gill morphology measurements (Dillon and
Goldstein, 1 984). The goal of PCA is to extract maximum
variance from the original data set with as few orthogonal
factors (components) as possible, thereby reducing the
variable to sample ratio and precluding statistical prob-
lems resulting from multicollinearity. Interpretations of
the derived principal components were based upon factor
loadings, which represent the correlations of the original
variables with the respective components (component-
variable correlations). Because loadings with the largest
absolute magnitudes have the greatest influence on the
components, the subsequent description of each principal
component was based upon an appraisal of similarities
among those variables with the highest loadings on a given
component. We used factor (component) scores (estimates
of each sample's value on the derived components based
upon weighted combinations of its values on the original
variables) in place of the original gill morphology mea-
surements as dependent variables in comparing the mor-
phological features of marsupial and non-marsupial gills
throughout the collection period using analysis of variance
(ANOVA) (Tabachnick and Fidell. 1989) and Dunn's
multiple comparison procedure to establish an experi-
ment-wise error rate of 0.05 (Kirk, 1982).
Preparation of gills for scanning electron microscopv
(SEM)
Dissected gill specimens for scanning electron micros-
copy were fixed in 2% glutaraldehyde in 0.2 A/Sorenson's
sodium phosphate buffer (pH 7.2) at 4°C for 2 h, post-
fixed in 2% cacodylate buffered (pH 7.4) osmium tetroxide
for an hour, and rinsed with several changes of buffer. A
vibratome (Lancer Model 1000) was used to section the
gill specimens into 2-8 mm thick segments that were ei-
ther parallel or perpendicular to the dorsal-ventral axis.
This procedure exposed the frontal surface of the gill la-
mellae and allowed us to examine the arrangement and
morphology of the water tubes and the position of the
glochidia in the brood chambers. Specimens were later
dehydrated through a graded ethanol series, dried in a
Pelco CPD-2 critical point drier, mounted on aluminum
SEM stubs, and sputter-coated with gold-palladium (Pelco
Model SC-4). External gill features, primary and second-
ary water tubes, brood chambers, and lamellar tissues ex-
posed by sectioning were examined and photographed
with a Philips 5 1 5 scanning electron microscope operating
at 15 kV.
Three-dimensional reconstruction and water tube
volume calculation
We used a computerized 3D-reconstruction program
(PC3D, Jandel Scientific) to examine and quantify the
volumetric changes that take place in the water tubes and
brood chambers of marsupial gills as a consequence of
brooding. Tissues were prepared for light microscopy as
described above, and serial frontal cross sections (10 /urn
thick) of non-marsupial (inner demibranchs of males or
females) and marsupial gills (female pre-brooding and
brooding outer demibranchs) were photographed. The
water tubes, brood chambers, filaments, and interlamellar
tissue of every fifth section were visually aligned (Gaunt
and Gaunt, 1978) and digitized using a Summagraphics
digitizer (25 digitized sections/sample; 4 samples/gill type)
attached to a Zenith Z-386SX computer. We created a
three-dimensional image of the gill sample by stacking
the sections using the PC3D software; the program's vol-
ume calculation subroutines were used to determine the
volumes of the primary and secondary water tubes. The
final reconstruction represented a slice through the gill
approximately 1.25 mm high and perpendicular to the
filaments. A Kruskal-Wallis one-way ANOVA (SYSTAT
Statistical Software: Wilkinson, 1990) and distribution free
multiple comparisons based on rank sums (Hollander and
Wolfe, 1973) were used to test for differences between
standardized water tube volume measurements.
Results
Description and morphometric analysis of marsupial
and non-marsupial gills
The nomenclature and terminology used to describe
the gills of A. cataracta in the present study are similar to
those of previous descriptions of lamellibranch gills by
Ridewood ( 1 903 ) and Ortmann (1911). Compared to non-
marsupial demibranchs, the marsupial gills of all female
mussels collected throughout the study were subdivided
by additional interlamellar septa, resulting in shorter (an-
terior-posterior axis) water tubes and a lower mean fila-
ment/septum ratio (15.2 vs. 48.9) (Table I). Water tubes
(brood chambers) containing larvae were swollen to more
than 30 times their original non-brooding width (medial-
lateral axis), producing nearly a 24-fold increase in cross-
sectional area and causing the ventral edge of the demi-
branch to expand into a thin, non-ciliated connection be-
tween opposing lamellae. Conversely, brooding had little
effect on the spacing and length (anterior-posterior dis-
tance) of the water tubes of marsupial gills (Table 1). The
lamellar tissue of marsupial gills was slightly thinner than
in non-marsupial gills, especially during periods of larval
incubation, but still possessed well-developed interfila-
ment water canals leading to ostial openings in the la-
GILL MORPHOLOGY OF ANODONTA CATAR.4CT.I
Table I
RCVH//S i mean ± SE) ofmorphometric analysis of male and female denubraneks during brooding and non-reproductive period^
149
Water tube
Collection period
Area
Width*
Length**
Gill
thickness
Lamellar
thickness
Filaments/
& gill type
(mnr)
(mm)
(mm)
(mm)
(mm)
Septum
n
Pre-brooding
Non-marsupial
0.158 ±0.021
0.193 ±0.017
1.028 ±0.078
0.974 ± 0.034
0.385 ±0.0 18
46.5 ±3.12
15
Marsupial
0.078 ±0.014
0. 1 35 ± 0.02 1
0.548 ±0.041
1.052 ±0.039
0.273 ± 0.027
16.0 ± 0.89
5
Gravid
Non-marsupial
0.193 ± 0.022
0.186 ±0.015
1.153 ±0.061
0.840 ± 0.033
0.333 ±0.0 12
46.2 ± 1.68
32
Marsupial
1.846 ±0.179
3.938 ± 0.205
0.457 ± 0.032
4.550 ±0.187
0.1 86 ±0.205
13.7 ±0.77
10
Secondary water tube
0.020 ± 0.005
0.063 ±0.010
0.370 ± 0.026
Post-release
Non-marsupial
0.233 ± 0.032
0.267 ± 0.032
1.077 ±0.055
1.130 ±0.076
0.391 ±0.032
51.1 ±2.46
15
Marsupial
0.185 ±0.050
0.768 ±0.169
0.298 ± 0.097
1.589 ±0.075
0.240 ±0.169
16.0 ±0.86
5
Secondary water tube
0.011 ±0.002
0.082 ±0.010
0.215 ±0.032
* Medial-Lateral axis.
** Anterior-Posterior axis.
Because there were no significant differences between the respective morphometric characters of all non-marsupial demibranchs. the data for the
inner demibranchs of females, and the inner and outer demibranchs of males, for each collection period have been pooled.
mellar walls of the secondary water tubes (Figs. 3, 4, 5).
Secondary septa were continuous with the interlamellar
septa (Figs. 3, 4) and lacked any apparent openings or
ostia leading from the secondary water tubes to the brood
chambers, effectively isolating the developing larvae from
water flowing through the mantle cavity and the secondary
water tubes (Fig. 6). Secondary septa also lacked the cil-
iated cells present on the lamellar walls of the primary
and secondary water tubes (compare Figs. 5. 6).
The interlamellar septa of non-marsupial gills were
continuous with the lamellar tissue and contained well-
developed arterial vessels (Fig. 7). Comparable vessels were
rarer in marsupial gills and were positioned at the base
of the septum near the junction with the lamellar tissue.
Figures 3 & 4. Frontal section of a marsupial demibranch (only one lamella is shown) containing mature
glochidia (G). Brood chambers (BC) are separated by thin interlamellar septa (ILS) which connect the
ascending and descending sides of the demibranch. Figure 4 is a higher magnification view of the highlighted
area in Figure 3 showing the position of the secondary septa (SS) forming the secondary water tubes (SWT).
Well-developed interfilament canals (IFC) are located between adjacent filaments (F) and lead to ostial
openings in the lamellar tissue (walls) of the water tubes. Nerves (N) located in the lamellar tissue are also
present.
150
R. A. TANKERSLEY AND R. V. DIMOCK, JR.
•
Figure 5. Scanning electron micrograph of the inner lamellar surface
of a secondary water tube of a gravid marsupial gill ofAnodonta cataracta.
Water pumped through the internlament canals by lateral cilia enters
the secondary water tubes through well-defined ostia (O). Lamellar walls
consisted primarily of ciliated epithelial cells (CE). The edges of the in-
terlamellar septa (ILS) connecting the opposing lamellae and forming
the anterior and posterior walls of the secondary water tube are also
visible.
Figure 6. Lateral view of the inner surface (brood chamber side) of
a secondary septum of a gravid marsupial demibranch (glochidia re-
moved). Note the absence of ciliated epithelial cells and ostia (see Fig.
5) leading from the secondary water tubes. Secondary septa are formed
by outgrowths of the interlamellar septa (ILS) prior to larval incubation.
Distended ovisacs and numerous secondary water tubes
were still present in the gills of females collected just after
glochidial release (Fig. 8), but were not present in gills
prior to brooding (Fig. 9). Brooding had no apparent effect
on the presence or distribution of frontal cilia or latero-
frontal cirri, which were visible on the surface of the gill,
and all gills lacked the frontal cirri recently reported found
in some other freshwater species (Way el al., 1989).
Moreover, there were no seasonal differences between
males and females in the morphology of their non-brood-
ing demibranchs.
Principal component analysis on the six gill morphol-
ogy variables resulted in two components (PCI & PC2)
being retained (eigenvalues > 1 ) that explained approxi-
mately 89% (PCI = 49%; PC2 = 40%) of the total variance.
An orthogonal rotation (varimax) was performed on the
extracted factors (components) to improve their inter-
pretability while still maintaining independent factor
scores. Water tube area and width and gill thickness all
had high loadings on PCI and represented morphological
features associated with the gill's left-right axis dimension
(Fig. 10). The remaining three variables, water tube length,
filaments/septum ratio, and lamellar tissue thickness, had
high loadings on PC2 and characterize features associated
with the arrangement, spacing, and number of water tubes.
Therefore, PCI and PC2 were respectively labeled "left-
right axis thickness" and "water-tube compactness."
The mean factor scores for both PC 1 and PC2 for each
type of gill are plotted in Figure 1 1. There were no sig-
nificant differences between the inner and outer demi-
branchs of males (PCI: F = 5.03; PC2: F = 4.58; d.f.
= 1, 20; P > 0.05) or between the inner demibranchs of
males and females (PCI: F = 7.63; PC2: F = 4.83; d.f.
= 1, 39; P > 0.05). Because there were no seasonal dif-
ferences in the factor scores of any of these non-marsupial
gills (PCI: F= 1.28;PC2 = 1.36; d.f. = 2, 59, P > 0.05),
all scores for each type of gill were pooled to simplify the
analysis. Marsupial gills containing larvae had higher PCI
factor scores than pre- or post-brooding gills (F = 81.40;
d.f. = 2, 17; P < 0.01). The arrangement of the water
tubes (PC2) of marsupial gills differed significantly from
that of non-marsupial gills (F = 70.22; d.f. = 1, 36; P
< 0.01) but remained consistent throughout the collection
period (i.e., exhibited no significant seasonal variation; F
= 2.57; d.f. = 2, 17;/>>0.05).
3D reconstructions and comparison of primary and
secondary water tube volumes
The mean volumes of 1-mm sections of each type of
water tube are listed in Table II. Because the total number
of tubes present in each demibranch varied with the type
of gill (marsupial or non-marsupial) and the mussel's re-
productive condition (i.e., marsupial gills had two sec-
ondary water tubes/septum during brooding periods), the
volume measurements are also expressed as the volume
of water tube/ 100 gill filaments (counted as filaments
present on both the ascending and descending lamellae).
Although the volume of the primary water tubes of non-
marsupial gills, expressed as ml/mm of gill tissue, was
significantly larger than that of either the primary or sec-
ondary canals of marsupial gills (H = 9.85; d.f. = 2; P
GILL MORPHOLOGY OF ANODONTA CATAR.4CTA
151
Figure 7. Frontal view of the lamellar surface of a non-marsupial
demibranch. Water pumped by lateral cilia present on the gill filaments
(F) enters the interfilament canals (IFC), which empty into the primary
water tubes (PWT). Arterial vessels (AV) are located between water tubes
in the interlamellar septa (ILS). Nerves (N) situated in the lamellar tissue
are also visible.
Figure 8. Frontal section of a distended post-brooding marsupial
demibranch showing the empty brood chambers (BC) and the stretched
interlamellar septa (ILS). Most brood chambers still possessed well-de-
veloped secondary septa (SS) forming secondary water tubes on both the
lateral and medial sides of the gill, but some secondary septa are lost by
this post reproductive stage.
Figure 9. Representative section through a non-gravid marsupial
gill showing the crowded organization of the water tubes (PWT). Fol-
lowing larval release, secondary septa and water tubes disappear, the
< 0.01), differences in water tube volumes standardized
by filament number were not significant (H = 4.88; d.f.
= 2; P = 0.09). These data suggest that, although the
secondary water tubes of marsupial gills were significantly
smaller than the primary water tubes of non-marsupial
gills, the tripartite arrangement of the brooding demi-
branchs partially compensated for the blockage of the
brood chambers by developing larvae by supplying ap-
proximately the same total volume for irrigation.
Discussion
Ortmann's (1911) early descriptions of the anatomical
features of the gills ofAnodonta emphasized both the tri-
partite morphology of the marsupia during reproductive
periods, and the compact arrangement of their primary
water tubes, relative to the inner demibranchs of females
and all four demibranchs of males. The permanent dif-
ferentiation in the architecture of the marsupial gills of
A. cataracta is represented by the second principal com-
ponent (PC2) in the current study and remains one of the
few sexually dimorphic features of this species. The ma-
jority of changes in gill morphology associated with larval
incubation occurred within the left-right axis (represented
by PCI) but were only transient changes associated spe-
cifically with brooding. The unidirectional swelling of the
outer gill was accompanied by comparable alterations in
the size of the water tube walls and interlamellar septa,
but no changes in the spacing or arrangement of the fil-
aments. Furthermore, the presence, in gravid marsupial
gills, of well-developed interfilament water canals leading
to the secondary water tubes suggests that the marsupial
gills continue to transport water and presumably filter
particles despite striking changes in their morphology.
Brooding caused no corresponding changes in female
medial gills, such as an increase in water tube area, that
might offset changes in the lateral marsupial demibranchs.
Overall, the morphometric measurements reported in
the present study probably represent conservative esti-
mates of gill alterations associated with brooding, because
sample preparation, including fixation and dehydration,
caused some shrinkage of tissue (Humason, 1979; Gabriel,
1982). Furthermore, water tube measurements may only
approximate in vivo conditions, because pressure differ-
ences maintained by the cilia as they pump water between
the mantle cavity and the lumen of the demibranch cause
the demibranchs and water tubes to be inflated compared
to newly excised tissue (J0rgensen et al., 1986).
interlamellar septa (ILS) become thickened, and the interfilament water
canals (IFC) channel water through ostia located in the walls of the pri-
mary water tubes.
152
R. A. TANKERSLEY AND R. V. DIMOCK, JR.
Length (
Filaments/Septum •
Lamellar Tissue •
Thickness
PC 2
0.5
-0.5
PC 1
0.5
Width
-0.5
•Area
Gill Thickness
Figure 10. Pairwise plot of the factor loadings (PCI & PC2) for the
six morphometnc variables following orthogonal (varimax) rotation.
Many taxonomic schemes established for unionid
mussels (for example Ortmann. 1911, and Heard and
Guckert, 1971) rely heavily upon reproductive charac-
teristics associated with the marsupial demibranchs, in-
cluding the number (2 or 4) and location (inner or outer)
of the marsupia, the proportion of the demibranch used
for brooding, the location of developing larvae within the
gills, the arrangement of the brood chambers including
the presence of secondary water canals, the magnitude of
swelling of the lamellae, and the duration of the incubation
period. The bradytictic, ectobranchous, tripartite mar-
supial arrangement of anodontine mussels, including A.
cataracta, is considered to be more specialized ("ad-
vanced") than that of other mussels, which characteris-
tically have shorter incubation periods, more marsupial
demibranchs (tetragenous), and no secondary water tubes
( Ortmann, 1911; Heard and Guckert, 1971). The tripartite
arrangement of the marsupia of anodontine mussels has
long been linked to the lengthy incubation phase of their
breeding cycle, because it permits isolation of larvae while
providing passageways for the maintenance of water
transport. However, the assumption by most researchers,
including most recently Richard ct al. (1991), that sec-
ondary water tubes serve as lumina for irrigation during
brooding periods has not been confirmed.
Construction of temporary secondary septa and a thin
membrane at the dorsal end of the ovisacs of the lateral
gills of female A. cataracta provides formal barriers to the
circulation of water from the mantle cavity through the
marsupial gills and effectively isolates and protects the
larvae from the surrounding medium. Larval isolation
has also been documented in unionids that lack secondary
septa, including members of the Lampsilinae, and is
thought to be accomplished by the contraction of the ostia
that lead from the interfilamentar canals which, in turn,
restricts the flow of water into the water tubes (brood
chambers) (Richard et al., 1991). Transport of water in
these species is probably limited to only the demibranchs
or portions of demibranchs not containing glochidia. The
mechanism by which tetragenous mussels sustain water
transport during brooding is less obvious, because all four
demibranchs are used for larval incubation and retain the
marsupial morphology. In lampsilines, only a portion of
the gill is used for brooding; the remainder possesses water
tubes that are similar to those of non-marsupial demi-
branchs. This may be an alternative mechanism for sat-
isfying the conflicting demands of water transport required
for filtration, respiration, and larval incubation, and more
specialized than the tripartite arrangement of A cataracta
(Heard and Guckert, 1971; Kat, 1984).
In addition to serving as shelters for larval development,
brood chambers and the glochidial isolation they provide
might also facilitate the transfer of nutrients from the fe-
male to the developing larvae. As reported for other
unionids (Heard, 1975; Silverman et al.. 1987; Richard
et al.. 1991), the epithelia of the secondary and interla-
mellar septa of A. cataracta lack openings or ostia leading
to the secondary water canals and thereby limit direct
nutrient or ion exchange with external pond water. In-
vestigations of the maternal investment in larval nutrition
and development have been restricted primarily to ex-
aminations of the use of maternal calcium reserves for
the formation of larval shells (Silverman et al.. 1985,
1987). Although the mobilization of calcium concretions
in the gills of females and their subsequent incorporation
in the shells of brooded embryos is well documented, the
mechanism of transfer is still unknown. Wood ( 1974) re-
PC 2
Outer Pre- Brood ing
o
Outer: Release
PC 1
O
Outer Brooding
Figure 11. Pairwise plot of the mean factor scores (see Fig. 10: PCI
describes changes in the lateral-medial axis of the demibranchs and brood
chambers; PC2 describes the number and arrangement of water tubes
and filaments) for male and female inner and outer demibranchs
throughout the study. Because there were no significant differences in
the factor scores of female inner demibranchs and male inner and outer
demibranchs among the collection periods, the scores for each type of
gill were pooled to simplify the analysis.
GILL MORPHOLOGY OF ANODONTA CATARACT A
Table II
153
Water lube volume calculations (mean ± SE: n = 4) based upon 3D reconstructions of non-marsupial and marsupial gills of Anodonta cataracta
during brooding and pre-brooding (non-gravid) periods
Marsupial gill
Non-marsupial gill
PWT
Pre-brooding
Brooding
PWT
BC SWT
ml/mm of gill (Xl(T5)
ml/mm/100 filaments (xl(T5)
17.3 ±2.0
35.8 ± 4.2
5.3 ± 0.8
35.6 ± 5.4
161.1 ± 23.6 1.5 ±0.18
1090 ± 158 20.5 ±2.39*
* Includes two secondary water tubes per brood chamber.
Abbreviations: PWT. primary water tube; BC. brood chamber; SWT, secondary water tube.
ported that gravid Anodonta cygnea that were fed 14C-
lahelled algae incorporated significant concentrations of
the label in both the glochidia and interlamellar septa;
Wood also suggested that nutrients could have been
transferred from the female to the developing larvae via
mucus secreted by cells located in the interlamellar septa.
Although the lateral swelling of marsupial gills and the
packing of the brood chambers with larvae may isolate
many of the glochidia from direct contact with the lateral
and medial inner surfaces of the gill, the narrow (anterior-
posterior dimension) arrangement of the individual brood
chambers of A. cataracta keeps the larvae in close contact
with the interfilamentar septa and may facilitate the
transfer of nutrients from the female to developing larvae.
Gill irrigation and suspension feeding in bivalves are
dominated by the viscous forces characteristic of low
Reynolds numbers (< 1 ), producing a laminar flow of wa-
ter through the demibranchs (J0rgensen, 1982, 1983). The
pump, generated by the beating of lateral cilia present on
the gill filaments near the entrance to the interfilament
water canals, is influenced by, among other parameters,
the velocity of water passing through the interfilament
water canals and pressure heads produced by the lateral
cilia (Jorgensen el ai, 1988). Jorgensen et al. (1986) es-
tablished the following equation for analysis of pump and
system characteristics:
AHP = AH12 + AHr + AHex + AH,r
where AHP = pressure difference produced by the pump,
AH 1 2 = back pressure, AHf = frictional resistance of the
canal system (including the internlamentary water canals,
the water tubes and exhalent siphon), AHex = exit loss
generated by the kinetic energy of the water leaving the
exhalent siphon, and AH(f = active resistance produced
by the beating of the latero-frontal cirri located on the gill
filaments. Indirect estimates of the various components
of the ciliary pump of the marine mussel Alytilux echilis
revealed that interfilament canals constituted about 32%
of the total resistance in the system (Jorgensen et al., 1986),
but the frictional resistance produced by the lumen of the
water tubes was assumed to have only a negligible effect
on the pump. Although the morphology of the eulamel-
libranch gill of A. cataracta differs markedly from the
filibranch gill ofMytilus, and detailed comparisons of the
properties and energetics of the ciliary pumps of both types
of gills are unavailable, larval incubation may have a
greater impact on the molluscan pump of brooding eu-
lamellibranchs than is indicated by the current simplified
model based on non-brooding, filibranch bivalve char-
acteristics.
Because the flow resistance of a fluid passing through
a cylinder is extremely sensitive to reductions in bore size
( Vogel, 198 1 ), the use of a series of smaller diameter tubes
(secondary water tubes vs. primary water tubes) by gravid
marsupial gills most likely has a significant impact on the
resistance to flow and the cost of pumping. If the primary
and secondary water tubes are treated as a series of parallel
cylinders, according to Poiseuille's equation, volume flow
rate would vary with the fourth power of the tube's radius.
Therefore, even if the combined cross-sectional area of
the two secondary water tubes were equal to that of a
single primary water tube during non-reproductive peri-
ods, the overall flow rate in the two smaller tubes would
be only one-half that of the larger tube for a given pressure
change (Vogel, 1981). Consequently, estimates of flow
rates through secondary water tubes based upon the pres-
ent volume calculations (Table II) would predict that the
total flow in marsupial gills during brooding would only
be approximately 16% and 4% of that in primary water
tubes of non-gravid marsupial and non-marsupial gills,
respectively.
Maintenance of flow rates through the secondary water
tubes that are comparable to those through unobstructed
primary water tubes also would be energetically costly.
Because the power required to generate flow through a
cylinder is inversely related to the square of its radius, the
pressures that would be required to irrigate the smaller
diameter secondary water tubes likely exceed the capa-
154
R. A. TANK.ERSLEY AND R. V. DIMOCK, JR.
bility of a ciliary pump (J0rgensen el a/., 1986). Thus, the
flow rates within brooding marsupial gills are probably
much lower than those produced within non-gravid mar-
supial gills or non-marsupial gills, even though the volume
of lumina available tor water transport is compensated
for by the construction of the secondary water tubes.
Changes in total gill volume (swelling during brooding)
that modify the flow dynamics within the mantle cavity
and additionally restrict water transport, may further re-
duce the effectiveness of the marsupial gill. It is unclear
whether the extensive reorganization of marsupial gill tis-
sue following larval release permits the demibranchs to
assume functional characteristics of non-marsupial gills
after the brooding season. Investigations are currently un-
derway in our laboratory to assess the impact of larval
incubation on the pumping and feeding physiology of A.
calaracta.
Acknowledgments
The current study was funded in part by a NSF Dis-
sertation Improvement Grant (BSR-9001345) and a
Theodore Roosevelt Memorial Grant from the American
Museum of Natural History to the senior author. We are
grateful to Dr. N. S. Allen for the use of the video mi-
croscopy facility and to P. E. Richard el al. for allowing
us to review a preprint of their manuscript. Special thanks
are due to J. Sizemore for sharing her microscopical ex-
pertise. J. Fernandez, E. Wetzel, and two anonymous re-
viewers provided valuable comments on the manuscript.
Literature Cited
Bloomer, H. H. 1934. On the sex, and sex-modification of the gill of
Anodonta cygnea. Proc. Malacol. Soc. Lend. 21: 21-28.
Coker, R. E., A. F. Shira, H. W. Clarke, and A. D. Howard. 1921.
Natural history and propagation of fresh-water mussels. Bull. U. S.
Bur Fish. 37: 75-181.
Dietz, T. H., and R. H. Alvarado. 1970. Ionic regulation in Lumbricus.
Bwi Bull. 141: 472-484.
Dillon, W. R., and M. Goldstein. 1984. Mullivariale Analysis. John
Wiley & Sons, Inc., New York. 587 pp.
Gabriel, B. L. 1982. Biological Scanning Electron Microscopy. Van
Nostrand Reinhold Co., New York. 186 pp.
Gaunt, P. N., and W. A. Gaunt. 1978. Three Dimensional Reconstruc-
tion in Biology. University Park Press, Baltimore. 1 74 pp.
Gordon, M. E., and D. G. Smith. 1990. Autumnal reproduction in
Cumberlandia nmnodonta (Unionoidea: Margaritifendae). Trans.
Am. Microsc. Soc. 109: 407-411.
Heard, W. H. 1975. Sexuality and other aspects of reproduction in
Anodonta (Pelecypoda: Unionidae). Malacologia 15: 81-103.
Heard, W. H., and R. H. Guckert. 1971. A re-evaluation of the Recent
Unionacea (Pelecypoda) of North America. Malacologia 10: 333-
355.
Hoeh, W. R. 1990. Phylogenetic relationships among eastern North
American Anodonla (Bivalvia: Unionidae). Malacol. Rev. 23: 63-
82.
Hollander, M., and D. A. Wolfe. 1973. Nonparametric Statistical
Methods. John Wiley & Sons. New York. 503 pp.
Humason, G. L. 1979. Animal Tissue Techniques. W. H. Freeman and
Co., San Francisco. 661 pp.
Jergensen, C. B. 1982. Fluid mechanics of the mussel gill: the lateral
cilia. Mar. Biol. 70:275-281.
Jergensen, C. B. 1983. Fluid mechanical aspects of suspension feeding.
Mar. Ecol. Prog. Ser. 11: 89-103.
Jergensen, C. B., P. S. Larsen, F. Mehlenberg, and H. U. Riisgard.
1988. The mussel pump: properties and modelling. Mar. Ecol. Prog.
Ser. 45:205-216.
Jergensen, C. B., P. Famme, H. S. Kristensen, P. S. Larsen, F. Meh-
lenberg, and H. U. Riisgard. 1986. The bivalve pump. Mar. Ecol.
Prog. Ser. 34: 69-77.
Kat, P. W. 1984. Parasitism and the Unionacea (Bivalvia). Biol. Rev.
59: 189-207.
Kays, W. T., H. Silverman, and T. H. Dietz. 1990. Water channels
and water canals in the gill of the freshwater mussel. Ligumia sub-
rostrata: ultrastructure and histochemistry. / Exp Zoo/. 254: 256-
269.
Kirk, R. E. 1982. Experimental Design. Brooks/Cole Publishing Co.,
California. 911 pp.
Lefevre, G., and W. C. Curtis. 1910. The marsupium of the Unionidae.
Biol. Bull. 19:31-34.
Moore, H. J. 1971. The structure of the latero-frontal cirri on the gills
of certain lamellibranch molluscs and their role in suspension feeding.
Mar. Biol. 11: 23-27.
Ortmann, A. E. 1911. A monograph of the najades of Pennsylvania.
Mem. Carnegie Mus. 4: 279-347.
Owen, G. 1974. Studies on the gill of Mytilus edulis: the eulatero-
frontal cirri. Proc. R Soc. Land. 187: 83-91.
Peck, R. H. 1877. The minute structure of the gills of the lamellibranch
Mollusca. Q J Microsc. Sci. 17: 43-66.
Richard, P. E., T. H. Dietz, and H. Silverman. 1991. Structure of the
gill during reproduction in the unionids, Anodonta grandm, Ligumia
subrostrala and Canincu/ma parva lexasensis. Can. J. Zoo/. 67: 1 744-
1754.
Ridewood, W. G. 1903. On the structure of the gills of Lamellibranchia.
Phil. Trans. Roy Soc. B 211: 147-284.
Silverman, H., W. L. Steffens, and T. H. Dietz. 1983. Calcium con-
cretions in the gills of a freshwater mussel serve as a calcium reservoir
during periods of hypoxia. J. Exp. Zoo/. 227: 177-189.
Silverman, H., W. L. Steffens, and T. H. Dietz. 1985. Calcium from
extracellular concretions in the gills of freshwater unionid mussels is
mobilized during reproduction. J Exp. Zoo/. 236: 137-147.
Silverman, H., W. T. Kays, and T. H. Dietz. 1987. Maternal calcium
contribution to glochidial shells in freshwater mussels (Eulamelli-
branchia: Unionidae). / Exp. Zoo/. 242: 137-146.
Silverman, H., P. E. Richard, R. H. Goddard, and T. H. Dietz. 1989.
Intracellular formation of calcium concretions by phagocytic cells in
freshwater mussels. Can J. Zoo/. 67: 198-207.
Silvester, N. R. 1988. Hydrodynamics of flow in Mytilus gills. J. Exp.
Mar. Biol. Ecol. 120: 171-182.
Silvester, N. R., and M. A. Sleigh. 1984. Hydrodynamic aspects of
particle capture by Mytilus. J. Mar. Biol. Assoc. U.K 64: 859-879.
Tabachnick, B. G., and L. S. Fidell. 1989. Using Multivanate Statistics.
Harper and Row, New York. 746 pp.
Vogel, S. 1981. Life In Moving Fluids. Princeton University Press,
New Jersey. 352 pp.
Way, C. M., D. J. Hombach, T. Deneka, and R. A. Whitehead. 1989. A
description of the ultrastructure of the gills of freshwater bivalves,
including a new structure, the frontal cirrus. Can. J. Zoo/. 67: 357-
362.
Wilkinson, L. 1990. SYSTAT: The System for Statistics. SYSTAT,
Inc. Evanston, IL. 676 pp.
Wood, E. M. 1974. Development and morphology of the glochidium
larva of Anodonta cygnea (Mollusca: Bivalvia). / Zoo/. 173: 1-13.
Reference: Biol. Bull 182: 155-158. (February, 1992)
Allorecognition in Colonial Marine Invertebrates:
Does Selection Favor Fusion with Kin
or Fusion with Self?
MICHAEL FELDGARDEN1 AND PHILIP O. YUND2
1 Department of Biology, Yale University, New Haven, Connecticut 06511 and2 Department
of Biological Sciences, University of New Orleans, New Orleans, Louisiana 70148
Previous analyses of the selective forces operating on al-
lorecognition systems in colonial marine invertebrates have
suggested that advantages to fusion with kin have selected
for the ability to recognize and fuse with related colonies.
While this explanation is compatible with the observation
of aggregated settlement of fusible larvae in an ascidian
species, it is not compatible with two other prominent fea-
tures ofallorecognition systems — the extensive allorecog-
nition allele polymorphism commonly observed in natural
populations and the recently reported instability of chimeric
colonies. We suggest that selection for fusion with self,
rather than fusion with kin, offers a more parsimonious
explanation for the two features listed above. Consequently,
self fusion may be a major selective force acting on allo-
recognition systems in colonial invertebrates.
Colonial marine invertebrates typically possess allore-
cognition systems that control fusion and rejection among
conspecific colonies, and such systems are either known
or expected to be genetically based ( 1 ). The broad distri-
bution of allorecognition in colonial taxa and the parallels
between invertebrate allorecognition and the vertebrate
immune system have spawned considerable interest in
the evolution ofallorecognition systems. Previous analyses
of the selective forces at work in allorecognition systems
have suggested that allorecognition mediates the costs and
benefits of fusion with conspecifics (1, 2). Fusion with
conspecifics is expected to confer benefits by increasing
the size of the resulting chimeric colony (1-3), which in
turn decreases the susceptibility of a colony to the impact
of ecological processes and increases colony reproductive
output (4-7). However, potential costs are incurred in
Received 28 June 1991; accepted 22 November 1991.
chimeric colonies as well. Because all genotypes in a chi-
mera have access to the production of gametes (4), one
genotype can functionally parasitize other members of
the chimera by contributing disproportionately to gamete
production (somatic cell parasitism) (4, 8). Allorecognition
systems are thus thought to function to limit fusion to
close relatives, so that the benefits of fusion can be ac-
quired while the potential costs are reduced (2). By re-
stricting fusion to close relatives, allorecognition systems
may constrain somatic parasitism to benefit a relative of
the victim, hence reducing the negative effect of parasitism
on a victim's fitness via a positive effect on inclusive fitness
(i.e., kin selection) (2).
Fusion with kin is a well-documented event, and hence
kin selection is certainly a potential force acting on allo-
recognition systems. However, we will argue that there is
an additional selective force at work, with the genetic in-
dividual as the target of selection, that is more likely to
explain two important features ofallorecognition systems.
Although our argument is applicable to most colonial taxa,
we will focus our discussion on botrylloid ascidians and
hydractiniid hydroids, the two groups for which the most
complete genetic and mechanistic data on allorecognition
are currently available.
At the heart of the kin selection argument are the an-
ticipated costs and benefits of fusion. We agree completely
with this assessment of costs and benefits, but suggest that
allorecognition plays a different role as well — to permit a
colony to obtain the benefits of fusion (increased colony
size) while completely eliminating the potential costs (so-
matic cell parasitism). We suggest that rather than simply
reducing the potential cost of somatic cell parasitism by
limiting fusion to close kin, selection favors colonies that
avoid these costs altogether by recognizing and fusing with
155
156
M. FELDGARDEN AND P. O. YUND
themselves (self fusion). Although fusion between kin oc-
curs, such events may simply represent mistakes in rec-
ognition due to the limitations of an imperfect system. If
kin fusion events are artifacts of allorecognition, kin se-
lection need not provide the dominant selective force
shaping allorecognition systems. Note that selection for
both self and kin fusion is generated by the potential for
somatic cell parasitism. While fusion with kin will limit
this cost, fusion with self will prevent it altogether.
Why would autogeneic fusion (fusion with the same
genotype; i.e., self fusion) be selectively favored? Frag-
mentation is an ubiquitous feature of colonial inverte-
brates, both through controlled fission and as a result of
extrinsic disturbance (9, 1 1 ). Surviving colony fragments
are likely to re-encounter their own genotype upon sub-
sequent growth and lateral expansion. If no mechanism
for self recognition exists, the subdivided colony will
compete with itself for space. Hence, autogeneic fusion
between colony fragments confers the advantages both of
an increase in size, as previously cited, and a release from
unnecessary competition.
While fragmentation and subsequent re-contact is likely
to be a major source of autogeneic fusion, self fusion is
common under other circumstances as well. Many co-
lonial invertebrates (including botrylloid ascidians and
hydractiniid hydroids) grow as epibionts on three-dimen-
sional substrata. Consequently, the growing margins of a
colony frequently encounter self upon wrapping around
the substratum. Even if physiological integration is main-
tained throughout the intervening regions of the colony,
the marginal tissue is confronted with a recognition prob-
lem. As in the case of fragmentation, these contacts impose
a recognize-or-compete constraint on the colony. In both
of these scenarios, autogeneic fusion confers the ecological
benefits effusion without the potential costs of allogeneic
fusion (fusion with other genotypes; generally kin fusion).
Early work on the population structure of corals and
sponges employed allorecognition as an assay of genetic
identity and hence explicitly assumed that fusion only
occurred among colonies derived from the same genotype
(i.e., that all fusion events were autogeneic) (12-14). These
studies were criticized for failing to provide adequate in-
dependent verification of genotypic identity (15), and fur-
ther work suggested that some fusion events were indeed
allogeneic ( 16, 17). The subsequent focus on the existence
of allogeneic fusion may have drawn attention from the
potential for autogeneic fusion in nature. While the ex-
plicit assumption of a correspondence between fusion and
genetic identity may not have been valid, the implicit as-
sumption that allorecognition systems served to permit
fusion with self may well have been accurate.
To this point, we hope that we have established that
the need to recognize and fuse with self theoretically pro-
vides a strong selective force on allorecognition systems.
In further evaluating the relative impact of selection for
kin and self fusion, we will discuss the compatibility of
these different selective forces with three documented fea-
tures of allorecognition systems in botrylloid ascidians
and hydractiniid hydroids — the extreme polymorphism
of allorecognition alleles, the instability of chimeras
formed by the fusion of related colonies, and the aggre-
gated settlement of fusible larvae.
High levels of allotype diversity are a prominent feature
of allorecognition systems in most colonial taxa (1). In
botrylloid ascidians, where the genetic mechanism of al-
lorecognition is known, allotype diversity is generated by
extensive allorecognition allele polymorphism, with the
number of alleles detected in natural populations ranging
from 40 to 100 (1, 1 1, 18-20). While the genetics of al-
lorecognition in hydractiniid hydroids has yet to be fully
resolved (2 1,22), fusion events and chimeric colonies ap-
pear to be very rare in natural populations (23), suggesting
a similarly high level of allele polymorphism.
Several authors have noted that kin selection does not
provide an obvious explanation for high allotype diversity
(1, 18, 24, 25). For high levels of polymorphism to occur,
alleles must increase in frequency when rare. However,
kin selection is likely to lead to the elimination of new
alleles, because rare alleles will be involved in few, if any,
fusion events (25). Consider the case of an hypothetical
new allele arising by mutation. This allele confers no im-
mediate advantage through kin selection, as kin fusion
events attributable to this allele are not possible. Selection
cannot favor the allele until it has already increased in
frequency to the point where kin fusion events are likely
to occur. Hence, any new allele is at a selective disadvan-
tage relative to established alleles and is likely to be elim-
inated. Population models based on the costs and benefits
of kin selection (though not explicitly modeling selection
at two levels) predict that the initially most frequent allele
will quickly increase to fixation in a population (25).
A closely related problem is that in order for the prem-
ises of kin selection to be valid, allotype must indicate the
degree of relatedness between colonies. However, the pos-
session of a given allorecognition allele is indicative of
relatedness only when that allele is rare (25). As an allele
increases in frequency, colonies that share this allele are
less likely to have inherited it from a recent common
ancestor. The sharing of a common allorecognition allele
is thus a very poor assay of relatedness. As an allele be-
comes common, fusion will occur between less related
genotypes, leading to the costs of somatic cell parasitism.
Due to the failure of alleles to increase when rare and to
convey accurate information on relatedness, the obser-
vation of high levels of polymorphism at allorecognition
loci in natural populations is difficult to reconcile with
kin selection.
SELECTION FOR SELF FUSION
157
The assumption of a selective advantage to self fusion
rather than kin fusion alters the selective regime on al-
lorecognition alleles to simple frequency-dependent se-
lection. Rare alleles are at a selective advantage due to a
greater potential to correctly identify self, while common
alleles are at a selective disadvantage due to the propensity
to incorrectly identify allogeneic colonies as self. These
recognition errors incur for the bearer colony the potential
costs of fusion (i.e., somatic cell parasitism). Consider
again a new allele arising by mutation. This allele is ini-
tially at a selective advantage, as it can be responsible only
for autogeneic fusion events and not for accidental allo-
geneic events. As the allele increases in frequency, the
incidence of allogeneic fusion will also increase, and se-
lection on the allele via somatic parasitism will operate
in the opposite direction, preventing the allele from
reaching fixation in the population.
As a consequence of frequency-dependent selection,
high levels of allele polymorphism are expected in pop-
ulations (26). In addition to preserving new alleles that
enter the population by chance, frequency-dependent se-
lection may indirectly increase the rate at which new alleles
are generated. Frequency-dependent selection could favor
alleles at modifier loci that generate new alleles at the
target locus by two different mechanisms. Modifier alleles
could increase either mutation rates in the gene coding
for allorecognition (27) or recombination rates within this
gene sequence (intragenic recombination) (28). Although
frequency-dependent selection has previously been rec-
ognized as the most likely explanation for high allorecog-
nition allele diversity ( 1 ), an ecological scenario that gen-
erates frequency-dependent selection has been lacking.
In the colonial ascidian Botryllus schlosseri, chimeras
formed by the fusion of two colonies of different genotype
(allogeneic fusion) eventually terminate in either the ab-
sorption of one colony by the other or the separation of
the two colonies, often accompanied by the death of one
(29-3 1 ). Although initially noted in cases where the fusing
colonies shared only one allorecognition allele, this result
also occurs when both allorecognition alleles are shared
(32). This phenomenon is not restricted to ascidians; a
high percentage of allogeneic chimeras of the colonial hy-
droid Hvdractinia symbiolongicarpus are also unstable
(33). The break-up of chimeras in this species is linked to
the onset of reproduction, further suggesting that recog-
nition decisions are impacted by the potential for somatic
cell parasitism (33).
The observation that most chimeras formed by kin fu-
sion are unstable and temporary is clearly not compatible
with a system dominated by kin selection. Few benefits
are likely to be received by the participating colonies dur-
ing a short association, and the ultimate result is generally
deleterious to at least one of the participating colonies
(29-31). In fact, the negative consequences of kin fusion
might be expected to select against the very existence of
an allorecognition system. Such a system would only be
maintained if selectively favored for other reasons; i.e.,
the selective advantages to self fusion.
Larvae of B. schlosseri settle in proximity to both larvae
and adult colonies with which they share an allorecog-
nition allele and hence can fuse upon subsequent growth
and contact (34). In contrast, H. symbiolongicarpus larvae
settle at random with respect to fusible adults and larvae
(35). Larval settlement as a function of future fusibility
is the sole observation that we are aware of that is con-
sistent with kin selection but not with the selective pressure
of self fusion. If fusion with relatives is beneficial and
favored by selection, then the ability to exploit this in-
formation at the time of settlement will also be favored.
Conversely, selection for self fusion will not generate se-
lection for recognition at the larval stage (except perhaps
to avoid settlement near fusible larvae and adults).
Hence, allorecognition systems in botrylloid ascidians
and hydractiniid hydroids demonstrate features that may
have been shaped by the selective pressures of both fusion
with kin and fusion with self. Because the aggregation of
fusible larvae is absent in hydractiniid hydroids, this family
appears to display mainly the effects of selection for self
fusion. However, data on botrylloid ascidians provide
support for both hypotheses. This is not a surprising result,
as these two selective forces are not mutually exclusive
and their relative impact may vary among taxa. A more
rigorous evaluation of these two hypotheses will require
further data on the costs, benefits, and, especially, the
relative frequency of self and kin fusion. For the moment,
we suggest that current data are at least as consistent with
the selective pressures generated by self fusion as they are
with kin selection.
The resolution of the issue of the selective forces op-
erating within allorecognition systems has additional im-
plications on our understanding of evolutionary processes.
Selection has come to be viewed as a force that operates
on many different biological levels, leading some to suggest
that the major features of biological organization are es-
tablished during transitions in levels of selection (8. 36).
In this light, kin selection has provided a justification for
the suggestion that allorecognition systems function to
mediate and control conflict between two different levels
of selection, the genotype and the colony (3). If selection
is found to occur predominately at the individual level,
then allorecognition should be viewed in a very different
light, as a system that attempts to restrict selection to a
single level by maintaining synonomy between genotype
and colony. A logical extension of this perspective suggests
that new biological systems may arise not only to accom-
modate and organize new levels of selection, but also to
hinder hierarchical expansion.
158
M. FELDGARDEN AND P. O. YUND
Acknowledgments
We thank Leo Buss, John Francis, Steve Gaines, Howie
Lasker, Pam O'Neil, and Buki Rinkevich for comments
on earlier versions of this manuscript. Financial support
was provided by the National Science Foundation (OCE-
89-00325 and OCE-9 1-960 14).
Literature Cited
1. Grosberg, R. K. 1988. The evolution of allorecognition specificity
in clonal invertebrates. Q. Rev. Biol. 63: 377-412.
2. Buss, L. W., and D. R. Green. 1985. Histoincompatibility in ver-
tebrates: the relict hypothesis. Dev Com;'. Immunol 9: 191-201.
3. Buss, L. W. 1990. Competition within and between encrusting
clonal invertebrates. Trends Ecol. Evol. 5: 352-356.
4. Buss, L. W. 1982. Somatic cell parasitism and the evolution of
somatic tissue compatibility. Proc. Natl. Acad. Sci. U.S.A. 79: 5337-
5341.
5. Jackson, J. B. C. 1985. Distribution and ecology of clonal and
aclonal benthic invertebrates. Pp. 297-356 in Population Biology
and Evolution of Clonal Organisms. }. B. C. Jackson, L. W. Buss,
and R. E. Cook, eds. Yale University Press, New Haven, CT.
6. Hughes, T. P., and J. B. C. Jackson. 1985. Population dynamics
and life histories of foliaceous corals. Ecol. Monogr. 55: 141-166.
7. Wulff, J. L. 1986. Variation in clone structure of fragmenting coral
reef sponges. Biol J Linn. Soc. 27: 31 1-330.
8. Buss, L. \V. 1987. The Evolution of Individuality. Princeton Press,
NJ.
9. McFadden, C. S. 1986. Colony fission increases particle capture
rates of a soft coral: advantages of being a small colony. J. E\p.
Mar. Biol. Ecol 103: 1-20.
10. Lasker, H. 1990. Clonal propagation and populations dynamics
of a gorgonian coral. Ecology 71: 1578-1589.
1 1. Karakashian, S., and R. Milkman. 1967. Colony fusion compat-
ibility types in Bolryllus sMosseri. Biol. Bull. 133: 473.
12. Hildemann, W. H., I. S. Johnston, and P. L. Jokiel. 1979.
Immunocompetence in the lowest metazoan phylum: transplantation
immunity in sponges. Science 204: 420-422.
1 3. Jokiel, P. L., W. H. Hildemann, and C. H. Bigger. 1982. Frequency
of intercolony graft acceptance or rejection as a measure of popu-
lation structure in the sponge Callyspongia diffusa. Mar. Biol. 71:
135-139.
1 4. Niegel, J. E., and J. C. Avise. 1983. Clonal diversity and population
structure in a reef-building coral, Acropora cervicornis: self-recog-
nition analysis and demographic interpretation. Evolution 37: 437-
453.
15. Curtis, A. S. G., J. Kerr, and N. Knowlton. 1982. Graft rejection
in sponges: genetic structure of accepting and rejecting populations.
Transplantation 30: 362-367.
16. Neigel, J. E., and G. P. Schmahl. 1984. Phenotypic variation within
histocompatibility-defined clones of marine sponges. Science 224:
413-415.
17. Neigel, J. E., and J. C. Avise. 1985. The precision of histocom-
patibility response in clonal recognition in tropical marine sponges.
Evolution 39: 724-732.
18. Tanaka, K., and H. Watanabe. 1973. Allogemc inhibition in a
compound ascidian, Botryllus primigenus Oka. I. Processes and fea-
tures of "non-fusion" reaction. Cell. Immunol. 7: 410-426.
19. Mukai, H., and H. Watanabe. 1975. Distribution of fusion in-
compatibility types in natural populations of the compound ascidian,
Botryllus primigenus. Proc. Jpn. Acad. 51: 44-47.
20. Scofield, V. L., J. M. Schlumpberger, L. A. West, and I. L. Weiss-
man. 1982. Protochordate allorecognition is controlled by a MHC-
like gene system. Nature 295: 499-502.
21. Hauenschild, V. C. 1954. Genetische und entwichlungphysiolo-
gische Intersuchungen uber Intersexualitat und Gewebevertragli-
chkeit bei Hydractinia cchinata Flem. Wilhelm Roux' Arch. En-
twicklungsmech. 147: 1-41.
22. Hauenschild, V. C. 1956. Uber die Vererbung einer Gewebever-
traglichkeits-eigenschaftt bei dem Hydroidpolypen Hydractinia
echinata. Z. Naturforsch. lib: 132-138.
23. Yund, P. O., and H. M. Parker. 1989. Population structure of
Hydractinia sp. nov. C in the Gulf of Maine. J. Exp. Mar. Biol.
Ecol. 125: 63-82.
24. Crozier, R. H. 1986. Genetic clonal recognition abilities in marine
invertebrates must be maintained by selection for something else.
Evolution 40: 1100-1101.
25. Grosberg, R. K., and J. F. Quinn. 1988. The evolution of allore-
cognition specificity. Pp. 157-167 in Invertebrate Historecognition.
R. K. Grosberg, D. Hedgecock, and K. Nelson, eds. Plenum Press,
New York.
26. Ayala, F. J., and C. A. Campbell. 1974. Frequency-dependent se-
lection. Ann. Rev. Ecol. Sysl. 5: 1 15-138.
27. Ohno, S. 1969. The spontaneous mutation rate revisited and the
possible principle of polymorphism generating more polymorphism.
Can. J. Genet. Cytol. 11: 457-467.
28. Bodmer, W. F., and A. J. Darlington. 1969. Linkage and recom-
bination at the molecular level. Pp. 223-265 in Genetic Organization.
Vol. 1. E. W. Caspari and A. W. Ravin, eds. Academic Press, New
York.
29. Rinkevich, B., and I. L. Weissman. 1987. A long-term study on
fused subclones in the ascidian Botryllus schlossen: the resorption
phenomenon (Protochordata: Tunicata). J. Zoo/ (Land.) 213: 717-
733.
30. Rinkevich, B., and I. L. Weissman. 1987. The fate of Bolryllus
larvae cosettled with parental colonies: benefical or deleterious con-
sequences? Biol. Bull. 173: 474-488.
3 1 . Rinkevich, B., and I. L. Weissman. 1989. Variation in the outcomes
following chimera formation in the colonial tunicate Botryllus
schlosseri. Bull. Mar. Sci. 45: 213-227.
32. Weissman, I. L., and B. Rinkevich. 1990. Allorecognition histo-
compatibility in a protochordate species: is the relationship to MHC
somatic or structural? Immunol. Rev 113: 227-241.
33. Shenk, M. A., and L. W. Buss. 1991. Ontogenetic changes in fu-
sibility in the colonial hydroid Hydractinia symbiolongicarpus. J
Exp. Zool. 257: 80-86.
34. Grosberg, R. K., and J. F. Quinn. 1986. The genetic control and
consequences of kin recognition by the larvae of a colonial marine
invertebrate. Nature 332: 456-459.
35. Yund, P. O., C. W. Cunningham, and L. W. Buss. 1987. Recruit-
ment and postrecruitment interactions in a colonial hydroid. Ecology
68:971-982.
36. Maynard Smith, J. 1988. Evolutionary progress and levels of se-
lection. Pp. 219-230 in Evolutionary Progress. M. H. Nitecki, ed.
University of Chicago Press. Chicago. 1L.
Reference: Biol. Bull 182: 159-162. (February, 1992)
Avoidance of Hypoxia in a Cnidarian Symbiosis
by Algal Photosynthetic Oxygen
M. L. RANDS. A. E. DOUGLAS*. B. C. LOUGHMAN, AND R. G. RATCLIFFE
Depart men! of Plant Sciences and * Department ofZoologv,
South Parks Road, Oxford 0X1 3RB, U.K.
The algal symbionts in a variety of invertebrates are
widely believed to increase the oxygen tension in the an-
imal tissue by producing photosynthetic oxygen ( 1 ). This
could be advantageous to the animal by maintaining nor-
moxia in anoxic waters (2-4). Equally, it could be detri-
mental, through hyperoxia and the generation of toxic
oxygen radicals (5), and this may contribute to the re-
current incidence of mass bleaching in tropical cnidarian
symbioses over the last decade (6). Here, we use in vivo
31P nuclear magnetic resonance spectroscopy (NMR) to
assess the effect of photosynthetic oxygen production by
the symbiotic alga Symbiodiniwn sp. on the energy me-
tabolism of a sea anemone, Anemonia viridis and, by using
acidification of the tissues and elevated ADP/ATP ratios
as linked indices of anaerobiosis (7), we show unequivo-
cally that photosynthetic oxygen can protect an inverte-
brate from hypoxia. Illumination prevents the rapid acid-
ification and reduction in ATP that occurs under hypoxic
conditions in the dark, and we suggest that this effect could
be partly responsible for algal enhancement of coral cal-
cification.
31P NMR spectra of Anemonia viridis tentacles showed
signals from a range of phosphorylated metabolites in-
cluding phosphonates, orthophosphate (P, ), and ATP (Fig.
1 ). The spectra were assigned on the basis of an earlier
31P NMR study of sea anemones (8) and from the char-
acteristic chemical shifts of the signals in the spectra. The
intensity of the ATP, ADP, and P, resonances, as well as
the position of the P, signal, were of particular interest
here, and the analysis of the spectra was confined to these
regions of the spectrum. The chemical shift of the P, signal
Received 30 August 1991; accepted 21 November 1991.
was measured relative to the narrow signal from a capillary
containing phosphocreatine; for the present purpose, the
accuracy with which this measurement could be made
outweighed any disadvantage arising from the overlap of
this reference signal with weak endogenous signals from
phosphocreatine and phosphoarginine.
Symbiotic tissue showed a high level of ATP when in-
cubated in the dark with a circulating oxygenated medium
(Fig. la). In contrast, when tissue was incubated in the
dark without the circulation of an oxygenated medium,
hypoxia developed, and the ATP level fell rapidly, be-
coming undetectable after 6 h (Fig. Ib). A number of
other metabolic changes, all characteristic of the transition
to the hypoxic state and well-documented in earlier in
v/vo31P NMR studies of animal tissue (9-1 1), were also
observed in the hypoxic tentacles. ADP, which was un-
detectable in the well-oxygenated tissue (Fig. la), became
measurable during the early stages of hypoxia (Fig. Ic),
and the increase in the ADP/ATP ratio to values greater
than 1 , could be followed for 3 h before the absolute levels
of both metabolites became unmeasurable. Breakdown
of the nucleotides caused a three-fold increase in the P,
level, and the shift in the position of the P, signal (see
below) reflected the expected acidification of the cyto-
plasm under hypoxic conditions. Illumination of the tissue
reversed all of these spectroscopic changes, and when light
was provided continuously, the absence of an external
oxygen supply had no effect on either the ATP level or
the cytoplasmic pH (Fig. Id). This striking result is con-
sistent with a number of earlier observations, including:
(i) assays of adenylates, which showed that ATP was re-
duced and ADP/ATP increased in the nonsymbiotic ane-
mone Bunodosoma cavernata during hypoxia (12); and
(ii) oxygen flux data, which suggested that the symbiotic
159
160
M. L. RANDS ET AL.
PCr
(Q)
20
0
-20 ppm
Figure 1. }t P NMR speclra of three different samples of Anemonia viridis ten-
tacles: (a) bathed in a circulating oxygenated medium in the dark: (b.c) under hypuxic
conditions, t e . without an external oxvgen supply, in the same medium in tin' dark:
and (d) under hypoxic conditions in the light The spectra were recorded at 121.49
MHz on a Bnikcr CXP300 NMR spectrometer using a selective frequency "P probe-
head. Approximately 50 tentacles from specimens collected from Portsmouth Harbour.
U.K.. were suspended m an artificial seawater medium (545 mM NaCI. 10 mM
KCI, 10 mM Tris. pH 8 3) and packed into a 10 mm NMR tube attached to an
experimental arrangement (2J) that allowed the suspending medium to be circulated
through the lube. Normoxic conditions were maintained with the oxygenated medium
/lowing at 10 ml mm'1: hypoxic conditions were achieved by slopping the flow and
allowing the tissue to deplete the oxygen in the small volume (approx. 9.5 ml) of
medium m the tube. The tissue was illuminated using a 150W Schott cold light
source (KL-1 SOOT) and a fiber optic cable. 1 m long, inserted through the probeltead
to the NMR tube This arrangement produced 40 iiE m~! s~' P.A R around the
sample in the probehead The t'ssite was maintained at 21-22°C and spectra were
anemone Anthopleura elegantissima can avoid oxygen
debt when placed under hypoxia in the light (4).
In vivo NMR extends our understanding of the oxygen
relations in Cnidaria by allowing changes in intracellular
pH and P, in response to hypoxia to be followed at the
same time as changes in ATP and ADP/ATP. The pH
dependence of the P, chemical shift allows cytoplasmic
pH values to be deduced in vivo from 31P NMR spectra
(13), and the intracellular pH in the tentacles was mea-
sured using a calibration curve based on the presumed
ionic composition of the cnidarian tissue (Fig. 2). In the
dark, with oxygen supplied by the circulating medium,
the average chemical shift for P, was 2.94 ± 0.02 ppm
(mean ± S.E. from 25 experiments) corresponding to an
intracellular pH of ~7.55 in line with the pH values re-
ported for other oxygenated marine invertebrates (9-
11,14-16). A similar value, 2.83 ± 0.01 ppm (n = 11),
corresponding to an intracellular pH of ~7.35, was ob-
tained in the light in the absence of an external oxygen
supply (Fig. 2), but in the dark, the pH fell rapidly, reach-
ing a value of 6.2 or less after 6 h of hypoxia. However,
when the hypoxic tissue was illuminated, the intracellular
pH returned to its normoxic value within 2 h (Fig. 2).
The degree of acidification in hypoxic tissues depends on
a number of factors (7), including the conditions under
which hypoxia develops (8), but the rapid acidification
observed here shows that Anemonia viridis has only weak
control over its intracellular pH during hypoxia.
It should be noted that the algal symbionts in A. viridis
represent less than 10% of the total biomass (17,18) and
that this was crucial to the success of the NMR experi-
ments for two reasons. First, there was insufficient algal
tissue to give detectable signals in the 3IP spectrum, al-
lowing the spectrum to be interpreted entirely in terms
of contributions from the host tissue. A similar conclusion
was reached in a study ofAiptasia pulcliellu (8) and it was
confirmed forAnenionia viridis by the negligible intensity
recorded in a dilute suspension of the isolated Symbio-
diniinn sp. (data not shown). Second, the low biomass of
the algal symbionts meant that the symbiotic tissue re-
quired very little light for photosynthesis, permitting a
relatively simple arrangement for illuminating the tissue
in the NMR tube. The 40 MEm'V P.A.R. provided by
the optical fiber should exceed the compensation point
recorded in 30-min blocks over a 6-h period using a 45° pulse angle and a 0.5-s
recycle lime. Spectrum (c) is a 30 min spectrum, recorded early in the hypoxic lime-
course, while the spectra in fa), (b), and (d) are sums of four 30-rnin blocks recorded
from 4 106 h after the start of each experiment. The resonance assignments are: 1,
phosphonates. 2. phosophomonoesters: 3. P,: 4. y-ATP: 5. a-ATP: 6. NAD(P)(H)
andNDP-hexo.se: 7. NDP-hexose: 8. p-ATP: PCr. phosphocreatine in the chemical
shift reference capillary Chemical shifts are quoted relative to the signal from 85%
onhophosphonc acid, but they were measured re/alive lo the signal at -2.44 ppm
from the PCr capillary
HYPOXIA AVOIDANCE THROUGH SYMBIOSIS
161
6
(ppm)
2-5 -\
2-0 -
1-5 -
-7-4 pH
-7-2
-7-0
-6-8
-66
-6-4
6-2
h6-0
Time (h)
Figure 2. Time dependence of the chemical shift (5) of the P, resonance for
Anemoma viridis tentacles in the absence of an external oxygen supply: (O) in the
light: l») in the dark: and (f>) initially in the dark followed by illumination- The
pH values were deduced from the chemical shift of the tissue P, resonance using a
calibration cuire obtained from a solution containing 190 mM A'C/, 100 mM NaCI.
0 1 >iM CaCI; . 3 mM Na:HPO, . 2 mM NaH,PO4 . The composition of this solution
was based on the presumed intracel/u/ar ionic composition ol cmdarian tissue (24-
26). and the calibration curve permitted pH measurements in the range pH 6-8
with an accuracy of±0 I pH units- Changes in pH could be detected with an acairacy
of ±0.05 pH units
for oxygen production in this system (19), and the spec-
troscopic changes observed on illumination are consistent
with the production of at least enough oxygen to meet
the respiration demands of the tentacles. In contrast, the
very high extinction coefficient of dense algal suspensions
and leaf tissues has. to a large extent, prevented the in-
vestigation of light-dark transitions in these tissues by
NMR (20). Thus the low biomass of the algal symbiont
favors the application of in vivo NMR and allows the
metabolic significance of the photosynthetically derived
oxygen to the host to be investigated directly.
Finally, these observations may have implications for
the mechanisms underlying calcification in cnidarian
corals. Recent respiratory experiments (21) have shown
that anthozoan polyps and colonies may actually be hyp-
oxic in the dark under well-stirred, air-saturated condi-
tions due to diffusion limitation of the oxygen supply.
Hypoxia inhibits calcium carbonate deposition (22), but
because we have shown that this condition can be pre-
vented in a symbiotic tissue by illumination, it may be
argued that algal symbionts promote calcification by pre-
venting hypoxia.
Acknowledgments
B. C. Loughman and R. G. Ratcliffe acknowledge the
financial support of the Agricultural and Food Research
Council. A. E. Douglas acknowledges the support of the
Royal Society of London, and M. L. Rands acknowledges
the receipt of a studentship from the Science and Engi-
neering Research Council.
Literature Cited
1 . Smith, D. C., and A. E. Douglas. 1987. The Biology of Symbiosis.
Edward Arnold, London. 302 pp.
2. Eaton, J. W., and J. O. Young. 1975. Studies on the symbiosis of
Phaenocora lyphlops (VEJDOVSKY) (Turbellaria; Neorhabdocoela)
and Chlorella vulgaris. Var. vulgaris. FOTT & NOVAKOVA
(Chlorococcales). Arch. Hydrobiol. 75: 50-75.
3. Boyle, J. E., and D. C. Smith. 1975. Biochemical interactions be-
tween the symbionts of Convohua roscoffensis. Proc. R Soc. Land.
B 189: 121-135.
4. Shick, J. M., and \V. I. Brown. 1977. Zooxanthellae-produced O2
promotes sea anemone expansion and eliminates oxygen debt under
environmental hypoxia. J. E\f>. Zoo/, 201: 149-155.
5. Dykens, J. A., and J. M. Shick. 1982. Oxygen production by en-
dosymbiotic algae controls superoxide dismutase activity in their
animal host. Suture 297: 579-580.
6. Lesser, M. P., W. R. Stochaj, D. W. Tapley, and J. M. Shick.
1990. Physiological mechanisms of bleaching in coral reef antho-
zoas: effect of irradience. ultra-violet radiation, and temperature on
the activities of protective enzymes against active oxygen. Coral Reefs
8: 225-232.
7. Hochachka, P. W., and T. P. Mommsen. 1983. Protons and an-
aerobiosis. Science 219: 1391-1397.
8. Steen, R. G. 1986. Impact of symbiotic algae on sea anemone
metabolism: analysis by in v/m>3'P Nuclear Magnetic Resonance
spectroscopy. / E.\p. Zoo/ 240: 315-325.
9. Kamp, G., and H. P. Juretschke. 1989. Hypercapnic and hypo-
capnic hypoxia in the lugworm Arenicola marina: a 3IP NMR study.
/ Exp. Zoo/. 252:219-227.
10. Raffin, J. P., M. T. Thebault, and J. Y. Legall. 1988. Changes in
phosphometabolites and intracellular pH in the tail muscle of the
prawn Pa/aemon serralus as shown by in vivo "P-NMR. / Comp.
Physio/. B 158: 223-228.
11 Wiseman, R. \V., \V. R. Ellington, and R. C. Rosanske. 1989.
Effects of extracellular pH and D-lactate efflux on regulation of
intracellular pH during isotonic contractions in a molluscan mussel:
a 3IP Nuclear Magnetic Resonance study. / Exp. Zoo!. 252: 228-
236.
12. Ellington, W. R. 1981. Effect of anoxia on the adenylates and the
energy charge in the sea anemone, Bunodosoma cavernata (Bosc).
Physio/. Zoo/. 54:415-422.
13. Gadian, D.G. 1982. Nuclear Magnetic Resonance and its Applications
to Living Systems. Clarendon Press, Oxford. 197pp.
14. Zange, J., \V. O. Portner, A. W. H. Jans, and M. K. Grieshaber.
1990. The intracellular pH of a molluscan smooth muscle during
a contraction-catch-relaxation cycle estimated by the distribution of
[14C] DMO and by "P-NMR spectroscopy. J. Exp. Biol 150: 81-
93.
15. Walsh, P. J., D. G. McDonald, and C. E. Booth. 1984. Acid-base
balance in the sea mussel, Mytilus edulis. II. Effects of hypoxia and
air-exposure on intracellular acid-base status. Afar. Biol. Lett. 5:
359-369.
16. Ellington, \V. R. 1983. The extent of intracellular acidification
during anoxia in the catch muscles of two bivalve molluscs. J. Exp.
Zoo/. 227: 313-317.
162
M. L. RANDS ET AL.
17. Tytler, E. M., and P. Spencer Davies. 1986. The budget of pho-
tosynthetically derived energy in the Anemonia sulcaia (Pennant)
symbiosis. J. Exp. Mar. Biol. Ecol. 99: 257-269.
18. Harland, A. D., L. M. Fixter, P. Spencer Davies, and R. A. Anderson.
1991. Distribution of lipids between the zooxanthellae and animal
compartment in the symbiotic sea anemone Anemonia viridis: wax
esters, triglycerides, and fatty acids. Mar. Biol. 110: 13-19.
19. Dorsett, D. A. 1984. Oxygen production in the intertidal anemone
Anemonia sulcala. Comp. Biochem. Physiol. 78A: 225-228.
20. Callies, R., R. Altenburger, A. Mayer, L. H. Grimme, and D.
Leibfritz. 1990. A new illumination system for /'/; vivo NMR spec-
troscopy. J. Magn. Reson. 90: 561-566.
21. Shick, J. M. 1990. Diffusion limitation and hyperoxic enhance-
ment of oxygen consumption in zooxanthellate sea anemones,
zoanthids, and corals. Biol. Bull. 179: 149-158.
22. Crisp, D. J. 1989. Tidally deposited bands in shells of barnacles
and molluscs. Pp. 103-124 in Origin, Evolution, and Modern Aspects
ofBiomineralization in Plants and Animals, R. E. Crick, ed. Plenum
Press, New York.
23. Lee, R. B., and R. G. Ratcliffe. 1983. Development of an aeration
system for use in plant tissue NMR experiments. J. Exp. Botany
34: 1213-1221.
24. Herrera, F. C, I. Lopez, R. Egea, and P. Zanders. 1989. Short-
term osmotic responses of cells and tissues of the sea anemone,
Condylactis gigantea. Comp. Biochem. Physiol. A. 92: 377-384.
25 Herrera, F. C., A. Rodriguez, I. Lopez, H. Weitzmann, and I. P.
Zanders. 1986. Characterization of cell ion exchange in the sea
anemone Condylactis gigantea. J. Comp. Physiol. B 156: 591-597.
26. Dunlap, K., K. Takeda, and P. Brehm. 1987. Activation of a cal-
cium-dependent photoprotein by chemical signalling through gap
junctions. Nature 325: 60-62.
The
Biological
Bulletin Board
February 1992
What's News at the BB...
•We now offer expedited processing for Research
Notes. We use express mail only, and offer a small emolu-
ment to reviewers to provide reviews within two weeks. A
carefully prepared Research Note will appear in print in as
little as 3.6 months after its submission. See our Research
Notes style guidelines outlined in our "Instructions to
Authors" section at the beginning of each issue.
• Authors who regularly send manuscripts of high quality to
the Bulletin will be considered Preferred Authors.
Preferred Authors benefit from expedited processing
for ALL of their submissions, and receive 200 free reprints
and color plates as needed. (Yes, we do color in the
Bulletin! See the Short Reports of the MBL General
Scientific Meetings in our October 1991 issue.)
•Back and special issues of the Bulletin are available:
— "The Woods Hole Marine Biological Laboratory," by
F. R. Lillie, an MBL Centennial supplement — $10
— "The Naples Zoological Station and The Marine
Biological Laboratory: One Hundred Years of
Biology," an MBL Centennial Supplement — $5
— "Consistency and Variability in Peptide Families,"
a symposium reprint from October 1989 — $9
— Regular single back issues, as available — $30 each
Contact our subscription manager, Christine Shoiwjlter, at our
Woods Hole office for further information.
Meeting Announcements
•The 20th Annual Marine Benthic Ecology Meeting will
be held in Newport, Rhode Island, on March 26 to 29,
1992. For further information contact Stanley Cobb,
Department of Zoology, University of Rhode Island,
Kingston, RI 02881; Tel: 401-792-2372.
•The International Conference on Molluscan
Conservation will be held at the University of Glasgow,
Scotland, on September 10 to 12, 1992. Sessions will
include taxonomy, distribution, legislation, and
conservation. For further information contact Fred
Woodward, International Conference on Molluscan
Conservation, Kelvingrove Museum & Art Gallery,
Kelvingrove, Glasgow G3 SAG, Great Britian;
Tel: (041) 357-3929.
•The international symposium on Climate Change and
Northern Fish Populations will be held in Victoria, British
Columbia, Canada, on October 13 to 16, 1992. Topics
include evidence for changes in climate and the resulting
effects in freshwater and marine environments; effects of
climate on fish populations; economic impacts of climate
change on fisheries; and preparing for climate change. For
further information contact the Symposium Secretary,
Department of Fisheries and Oceans, Pacific Biological
Station, Nanaimo, British Columbia, Canada V9R 5K6;
Tel: 604-756-7260.
CONTENTS
BEHAVIOR
Hermans, Colin ( )., and Richard A. Satterlie
Fast-strike feeding behavior in a pteropod mollusk,
Clione limacina Phipps
Wayne, Nancy L., and Gene D. Block
Effects of photoperiod and temperature on egg-lay-
ing behavior in a marine mollusk, Aplysia californica
DEVELOPMENT AND REPRODUCTION
Amemiya, S., and R. B. Emlet
The development and larval form of an echinothu-
rioid echinoid, Asthenosoma ijimai, revisited 15
Ausio, Juan
Purification and biochemical characterization of the
nuclear sperm-specific proteins of the bivalve mol-
lusks Agriodesma saxicola and Mytilimeria uuttalli .... 31
Blades- Eckelbarger, Pamela I., and Nancy H. Marcus
The origin of cortical vesicles and their role in egg
envelope formation in the "spiny" eggs of a calanoid
copepod, Centropages velifiratiu 41
Chandler, Resa M ., Mary Beth Thomas, and Julian
P. S. Smith, III
The role of shell granules and accessory cells in
eggshell formation in Convoluta pulchra (Turbellaria,
Acoela) 54
Chia, Fu-Shiang, Ron Koss, Shauna Stevens, and Jeff
I. Goldberg
Isolation of neurons of a nudibranch veliger .... 66
Holland, Linda /.., and Nicholas D. Holland
Early development in the lancelet (=amphioxus)
Branchiostoma floridae from sperm entry through
pronuclear fusion: presence of vegetal pole plasm
and lack of conspicuous ooplasmic segregation .. 77
Lee, Youn-Ho, and Victor D. Vacquier
The divergence of species-specific abalone sperm
lysins is promoted by positive Darwinian selection 97
ECOLOGY AND EVOLUTION
Gil-Turnes, M. Sofia, and William Fenical
Embryos of Homarus americanus are protected by
epibiotic bacteria 105
Williams-Howze, Judy, and Bruce C. Coull
Are temperature and photoperiod necessary cues
for encystment in the marine benthic harpacticoid
copepod Heteropsyllus nunni Coull? 109
GENERAL BIOLOGY
Jennings, Joseph B., Lester R. G. Cannon, and
Adrian J. Hick
The nature and origin of the epidermal scales of
Notodactylus handschini—an unusual temnocephalid
turbellarian ectosymbiotic on crayfish from north-
ern Queensland 117
Mangum, Charlotte P., James M. Colacino, and
Judith P. Grassle
Red blood cell oxygen binding in capitellid poly-
chaetes . 129
PHYSIOLOGY
Singarajah, K. V., and F. I. Harosi
Visual cells and pigments in a demersal fish, the
black sea bass (Centropristis striata) 135
Tankersley, Richard A., and Ronald V. Dimock, Jr.
Quantitative analysis of the structure and function
of the marsupial gills of the freshwater mussel An-
odonta cataracta 145
RESEARCH NOTES
Feldgarden, Michael, and Philip O. Yund
Allorecognition in colonial marine invertebrates:
does selection favor fusion with kin, or fusion with
self? 155
Rands, M. L., A. E. Douglas, B. C. Loughman, and
R. G. Ratcliffe
Avoidance of hypoxia in a cnidarian symbiosis by
algal photosynthetic oxygen 159
The Biological Bulletin Board 163
Volume 182
THE
Number 2
BIOLOGICAL
BULLETIN
APRIL, 1992
Published by the Marine Biological Laboratory
MBL 1992 Short Courses
MOLECULAR EVOLUTION August2 - August 14. 1992
APPLICATION DEADLINE: JUNE 1, 1992
A series of lectures and discussions exploring multiple approaches to molecular
evolution, and a computer laboratory for phylogenetic and sequence analysis. This
two week program is designed for a class of 60 established investigators, postdoctoral
fellows, and advanced graduate students. Director: Mitchell L. Sogin, Marine
Biological Laboratory.
METHODS IN COMPUTATIONAL NEUROSCIENCE
August 2 - August 29, 1992
APPLICATION DEADLINE: MAY 15, 1992
For 20 advanced graduate students and postdoctoral fellows in neurobiology,
physics, electrical engineering, computer science, and psychology. A background
in programming (preferably in C and UNIX) is highly desirable. This course
presents the basic techniques necessary to study single cells and neural networks
from a computational point of view and is organized around lectures, tutorials, and
computer laboratories. Directors: James M. Bower and Christof Koch, Computa-
tion and Neural System Program, California Institute of Technology.
FUNDAMENTAL ISSUES IN VISION RESEARCH: MOLECULAR
AND CELL BIOLOGICAL APPROACHES [sponsored by the National Eye
Institute. NIH] August 16 - August 29, 1992
APPLICATION DEADLINE: MAY 1, 1992
This laboratory-lecture course is intended for 20 graduate students and postdoctoral
fellows currently training in molecular biology, cell biology, and neurosciences
who are not currently involved in vision research. The goal of the course is to
present, in depth, the exciting theoretical and experimental approaches to funda-
mental research problem s in vision so that the students can evaluate the potential for
active research in this field. The faculty will describe and direct laboratories of on-
going research in the tissues of the eye of invertebrates and vertebrates. Costs of
attending the course, including travel, housing, and meals at MBL will be supported
by a scholarship fund from NEI. Directors: David S. Papermaster, University of
Texas Health Science Center, San Antonio; and John E. Dowling, Harvard
University.
RAPID MEASUREMENT OF NEUROTRANSMITTER SIGNALS
IN THE CENTRAL NERVOUS SYSTEM USING IN VIVO
ELECTROCHEMISTRY August 19 - August24, 1992
APPLICATION DEADLINE: JUNE 1, 1992
This course/workshop is intended for 16 graduate students, post-doctoral research-
ers, and investigators with interests in neuroscience, pharmacology, and chemo-
receptionwhowish to learn the art and practiceofmvivo electrochemistry as applied
to studies of the CNS. The course will address the theory and practice of these
techniques, and will encompass a combination of lectures, demonstrations, and
hands-on use of these methods and technologies. Directors: Greg A. Gerhardt and
Paul A. Moore, Rocky Mountain Center for Sensor Technology, University of
Colorado Health Sciences Center.
OPTICAL MICROSCOPY AND IMAGING IN THE BIOMEDICAL
SCIENCES September 23 - 30, 1992
APPLICATION DEADLINE: JULY 1, 1992
Designed primarily for 22 research scientists, physicians, postdoctoral trainees, and
advanced graduate students in animal, plant, and medical sciences as well as non-
biologists withexperience inmicroscopy and university faculty planning to develop
courses on similar topics. The course covers the fundamental theory and practical
use of modem optical microscopy. Special attention will be given to different
optical techniques and the newest photographic and video methods used in biological
and biomedical research. Directors: Nina SlrdmgrenAllen, Wake Forest University;
and Colin S. lizard, State University of New York at Albany.
For Further information and applkation forms, contact:
Florence Dwane, Admissions Coordinator • Office of Sponsored Programs ' '"- »
Marine Biotogkal Laboratory • Woods Hole, MA 02543, USA • (508) 548-3705, eiL 216
THE
BIOLOGICAL BULLETIN
PUBLISHED BY
THE MARINE BIOLOGICAL LABORATORY
Associate Editors
PETER A. V. ANDERSON, The Whitney Laboratory, University of Florida
DAVID EPEL, Hopkins Marine Station, Stanford University
J. MALCOLM SHICK, University of Maine, Orono
Editorial Board
DAPHNE GAIL FAUTIN, University of Kansas K. RANGA RAO, University of West Florida
WILLIAM F. GILLY. Hopkins Marine Station. STEVEN VOGEL, Duke University
Stanford University
Editor: MICHAEL J. GREENBERCi, The Whitney Laboratory, University of Florida
Managing Editor PAMELA L. CLAPP, Marine Biological Laboratory
APRIL, 1992
Printed and Issued by
LANCASTER PRESS, Inc.
PRINCE & LEMON STS.
LANCASTER, PA
THE BIOLOGICAL BULLETIN
THE BIOLOGICAL BULLETIN is published six times a year by the Marine Biological Laboratory, MBL
Street, Woods Hole, Massachusetts 02543.
Subscriptions and similar matter should be addressed to Subscription Manager, THE BIOLOGICAL BUL-
LETIN, Marine Biological Laboratory, Woods Hole, Massachusetts 02543. Single numbers, $30.00. Sub-
scription per volume (three issues), $77.50 ($155.00 per year for six issues).
Communications relative to manuscripts should be sent to Michael J. Greenberg. Editor-in-Chief, or
Pamela L. Clapp, Managing Editor, at the Marine Biological Laboratory, Woods Hole, Massachusetts 02543.
Telephone: (508) 548-3705, ext. 428. FAX: 508-540-6902. E-mail: pamcl@hoh.mbl.edu.
POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory,
Woods Hole, MA 02543.
Copyright © 1992, by the Marine Biological Laboratory
Second-class postage paid at Woods Hole, MA, and additional mailing offices.
ISSN 0006-3 185
INSTRUCTIONS TO AUTHORS
The Biological Bulletin accepts outstanding original research
reports of general interest to biologists throughout the world.
Papers are usually of intermediate length (10-40 manuscript
pages). A limited number of solicited review papers may be ac-
cepted after formal review. A paper will usually appear within
four months after its acceptance.
Very short, especially topical papers (less than 9 manuscript
pages including tables, figures, and bibliography) will be pub-
lished in a separate section entitled "Research Notes." A Re-
search Note in The Biological Bulletin follows the format of
similar notes in Nature. It should open with a summary para-
graph of 150 to 200 words comprising the introduction and the
conclusions. The rest of the text should continue on without
subheadings, and there should be no more than 30 references.
References should be referred to in the text by number, and
listed in the Literature Cited section in the order that they appear
in the text. Unlike references in Nature, references in the Re-
search Notes section should conform in punctuation and ar-
rangement to the style of recent issues of The Biological Bulletin.
Materials and Methods should be incorporated into appropriate
figure legends. See the article by Lohmann et al. (October 1990,
Vol. 179: 2 1 4-2 1 8) for sam'ple style. A Research Note will usually
appear within two months after its acceptance.
The Editorial Board requests that regular manuscripts con-
form to the requirements set below; those manuscripts that do
not conform will be returned to authors for correction before
review.
1. Manuscripts. Manuscripts, including figures, should
be submitted in triplicate. (Xerox1 copies of photographs are not
acceptable for review purposes.) The original manuscript must
be typed in no smaller than 12 pitch, using double spacing (in-
cluding figure legends, footnotes, bibliography, etc.) on one side
of 16- or 20-lb. bond paper, 8'/2 by 1 1 inches. Please, no right
justification. Manuscripts should be proofread carefully and er-
rors corrected legibly in black ink. Pages should be numbered
consecutively. Margins on all sides should be at least 1 inch (2.5
cm). Manuscripts should conform to the Council of Biology Ed-
itors Style Manual, 5th Edition (Council of Biology Editors,
1983) and to American spelling. Unusual abbreviations should
be kept to a minimum and should be spelled out on first reference
as well as defined in a footnote on the title page. Manuscripts
should be divided into the following components: Title page,
Abstract (of no more than 200 words). Introduction, Materials
and Methods. Results, Discussion, Acknowledgments, Literature
Cited, Tables, and Figure Legends. In addition, authors should
supply a list of words and phrases under which the article should
be indexed.
2. Title page. The title page consists of: a condensed title
or running head of no more than 35 letters and spaces, the
manuscript title, authors' names and appropriate addresses, and
footnotes listing present addresses, acknowledgments or contri-
bution numbers, and explanation of unusual abbreviations.
3. Figures. The dimensions of the printed page. 7 by 9
inches, should be kept in mind in preparing figures for publi-
cation. We recommend that figures be about 1 '/> times the linear
dimensions of the final printing desired, and that the ratio of
the largest to the smallest letter or number and of the thickest
to the thinnest line not exceed 1:1.5. Explanatory matter gen-
erally should be included in legends, although axes should always
be identified on the illustration itself. Figures should be prepared
for reproduction as either line cuts or halftones. Figures to be
reproduced as line cuts should be unmounted glossy photo-
graphic reproductions or drawn in black ink on white paper,
good-quality tracing cloth or plastic, or blue-lined coordinate
paper. Those to be reproduced as halftones should be mounted
on board, with both designating numbers or letters and scale
bars affixed directly to the figures. All figures should be numbered
in consecutive order, with no distinction between text and plate
figures. The author's name and an arrow indicating orientation
should appear on the reverse side of all figures.
4. Tables, footnotes, figure legends, etc. Authors should
follow the style in a recent issue of The Biological Bulletin in
preparing table headings, figure legends, and the like. Because
of the high cost of setting tabular material in type, authors are
asked to limit such material as much as possible. Tables, with
their headings and footnotes, should be typed on separate sheets,
numbered with consecutive Roman numerals, and placed after
the Literature Cited. Figure legends should contain enough in-
formation to make the figure intelligible separate from the text.
Legends should be typed double spaced, with consecutive Arabic
numbers, on a separate sheet at the end of the paper. Footnotes
should be limited to authors' current addresses, acknowledg-
ments or contribution numbers, and explanation of unusual
abbreviations. All such footnotes should appear on the title page.
Footnotes are not normally permitted in the body of the text.
5. Literature cited. In the text, literature should be cited
by the Harvard system, with papers by more than two authors
cited as Jones et al. 1980. Personal communications and ma-
terial in preparation or in press should be cited in the text only,
with author's initials and institutions, unless the material has
been formally accepted and a volume number can be supplied.
The list of references following the text should be headed Lit-
erature Cited, and must be typed double spaced on separate
pages, conforming in punctuation and arrangement to the style
of recent issues of The Biological Bulletin. Citations should in-
clude complete titles and inclusive pagination. Journal abbre-
viations should normally follow those of the U. S. A. Standards
Institute (USASI), as adopted by BIOLOGICAL ABSTRACTS and
CHEMICAL ABSTRACTS, with the minor differences set out below.
The most generally useful list of biological journal titles is that
published each year by BIOLOGICAL ABSTRACTS (BIOSIS List of
Serials; the most recent issue). Foreign authors, and others who
are accustomed to using THE WORLD LIST OF SCIENTIFIC PE-
RIODICALS, may find a booklet published by the Biological
Council of the U.K. (obtainable from the Institute of Biology,
41 Queen's Gate, London, S.W.7, England, U.K.) useful, since
it sets out the WORLD LIST abbreviations for most biological
journals with notes of the USASI abbreviations where these differ.
CHEMICAL ABSTRACTS publishes quarterly supplements of ad-
ditional abbreviations. The following points of reference style
for THE BIOLOGICAL BULLETIN differ from USASI (or modified
WORLD LIST) usage:
A. Journal abbreviations, and book titles, all underlined
(for italics)
B. All components of abbreviations with initial capitals
(not as European usage in WORLD LIST e.g., J. Cell. Comp.
Physiol. NOT J. cell. comp. Physiol.)
C. All abbreviated components must be followed by a pe-
riod, whole word components must not (i.e.. J. Cancer Res.)
D. Space between all components (e.g.. J. Cell. Comp.
Physiol., not J.Cell.Comp.Physiol.)
E. Unusual words in journal titles should be spelled out
in full, rather than employing new abbreviations invented by
the author. For example, use Rit V'isindafjelags Islendinga with-
out abbreviation.
F. All single word journal titles in full (e.g., Veliger. Ecol-
ogy, Brain).
G. The order of abbreviated components should be the
same as the word order of the complete title (i.e., Proc. and
Trans, placed where they appear, not transposed as in some
BIOLOGICAL ABSTRACTS listings).
H. A few well-known international journals in their pre-
ferred forms rather than WORLD LIST or USASI usage (e.g.,
Nature, Science. Evolution NOT Nature. Land., Science, N.Y.;
Evolution. Lancaster, Pa.)
6. Reprints, page proofs, and charges. Authors receive their
first 100 reprints (without covers) free of charge. Additional re-
prints may be ordered at time of publication and normally will
be delivered about two to three months after the issue date.
Authors (or delegates for foreign authors) will receive page proofs
of articles shortly before publication. They will be charged the
current cost of printers' time for corrections to these (other than
corrections of printers' or editors' errors). Other than these
charges for authors' alterations. The Biological Bulletin does not
have page charges.
iii
ERRATA
The Biological Bulletin, Volume 181, Number 3
Page 423, Table III
The following correction should be made in the article by R. K. Zimmer-Faust titled, "Chemical signal-
to-noise detection by spiny lobsters" (Biol. Bull. 181: 419-426):
On page 423, in Table III, the last entry in the W>column, which reads "3.65 (± 1.79) X 10~2" should
read, "3.65 (± 1.79) X 10~3." The exponent "~3" replaces the exponent "~2."
Page 4 27
The following correction should be made in the article by J. J. O'Brien el a/.. "Proteins of crustacean
exoskeletons: I. Similarities and differences among proteins of the four exoskeletal layers of four brachyurans"
(Biol. Bull. 181:427-441):
On page 427, the first footnote, which reads "Received" should read, "Received 24 April 1991; accepted
17 September 1991."
Page 499
The following correction should be made in the article by S. Soinila and G. J. Mpitsos titled, "Immuno-
histochemistry of diverging and converging neurotransmitter systems in mollusks" (Biol. Bull. 181:
484-499).
On page 499, the reference to Leonard el al., 1990, which reads "Leonard, J. L.. M. Martinez-Padron,
J. P. Edstrom, and K. Lukowiak. 1990. Does altering identified gill motor neuron activity alter gill behavior
in Aplysia? North Holland Publishing Co., Amsterdam" should read, "Leonard, J. L., M. Martinez-Padron,
J. P. Edstrom, and K. Lukowiak. 1990. Does altering identified gill motor neuron activity alter gill behavior
in Aplysia? Pp. 30-37 in Mollnscan Neurobiology. K. S. Kits. H. Boer, and J. Joose, eds. North Holland
Publishing Co., Amsterdam." The line, "Pp. 30-37 in Molluscan Neurobiology, K. S. Kits, H. Boer, and J.
Joose, eds.," should be added to the reference.
Reference: Biol Bull 182: 165-166. (April. 1992)
Carroll M.* Williams
DOROTHY M. SKINNER AND JOHN S. COOK
Biology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831
In spite of its history of keponed waters,
Virginia is loved by its sons and daughters.
From this land of ham and bourbon barrel
Emerged a Williams, christened Carroll.
A Southerner born, and a Richmond alumn.
He has lived in Cambridge since Kingdom Come.
But like natives of Hong Kong or even Tashkent,
He always retained his partic'lar accent.
In his first stint at Harvard, his doctor's degree
Crowned him Philos'pher of Biology.
He then mastered the art of prescription concoctor
And achieved the degree of a Medical Doctor.
But his expertise with scalpel and suture
Was devoted to insects alone in the future.
All matters insectuous engaged his attention.
From the many, a highlight or two shall we mention.
For example, consider the dry cleaner's plight.
Beleaguered, while cleaning, with insects in flight.
When the kelp flies bugged (or the kelp bugs flew),
The cleaners were frantic, but Carroll knew:
What was driving these people to grief and distraction
Was simply their own cleaning fluid's attraction.
* M stands for Milton. Muriel, Massachusetts, moth, Manduca. molt-
ing, metamorphosis, mitochondria, muscle, midge, magician, and much
more.
Carroll M. Williams died on October 11, 1991. He was one of America's
premiere insect physiologists, and he spent his entire career at Harvard
University. Over a period of forty years. The Biological Bulletin published
more than 25 of his seminal papers, and has chosen to honor the memory
of this extraordinary scientist with a poetic synopsis of some of the high-
lights of his career. This poem was delivered by Dorothy Skinner at the
banquet that concluded a symposium dedicated to Williams at the 1 980
annual meeting of the American Society of Zoologists. Carroll was en-
raptured by the rendition, and we hope it will convey to its readers some
of his spirit and creativity. Scholarly tributes to the career of Dr. Williams
will appear in Developmental Biology and the Biographical Memoirs of
the National Academy oj Sciences.
He was later enchanted, but not for too long,
By the siren-sweet sound of Drosophila 's song.
He made the humble moth Cecropia
A scientific cornucopia.
The careful perusal of hormonal data
Called his attention to corpora allata.
Switched on by the cold, just as Carroll was hopin'.
They discharged their prothoracicotropin.
Another Stoff with a different style
Is Cecropia's hormone juvenile.
To its own secretion it cannot adapt;
The pupa remains as a pupa entrapped.
This leads to a concept that's really quite scary:
Physiological hari kari.
165
166
D. M. SKINNER AND J. S. COOK
A woody factor that development fouls
He obtained from a column of paper towels.
A cost-cutting process was therein conceived;
With the factor eluted, the towels were retrieved.
Are corners required for Cecropia cocoons?
Carroll reared the poor beasts inside safe-sex balloons.
With the hypnotic charm of a true Southern preacher,
Carroll was born-to-the-art as a teacher.
He offered a course in Insect Disease.
The prerequisite was: "Can you read Japanese?"
Hungry students with stipends as teaching assistants
Were often hard-pressed just to manage subsistence.
But many were able to keep tummies quiet
With Carroll's tea cakes as their principal diet.
An integral part of his craft professorial
Is the camera work of his cohort Muriel.
In his published work, among the fixtures
Supporting the data are Muriel's pictures.
Surely there can be no finer alliance
Than elegant art wed to elegant science.
We close these lines of appreciation
With a word or two on sub-speciation.
A genetic analogy may not be prudent,
But he transmitted something to each Williams student.
For six-legged studies he set new criteria;
With his standards so high, we disdain the inferia.
Reference: Bial. Bull 182: 167-168. (April, 1992)
How the Axon Got its Tale
(For K.D.)
DEFOREST MELLON, JR.
Department of Biology, University of Virginia, Charlottesville, Virginia 22901
In early days the axon was unknown and quintessential
Till Bernstein's brilliant insights gave its secrets new
potential.
With pluses out and minuses within the membrane's con-
fines.
There must exist a gradient of concentrated ions.
An impulse in its travels down an axon would exist
Because ionic permeances undergo a twist.
Potassium is high within and sodium without.
And transient membrane breakdown then would foster
turnabout.
Short circuit currents would ensue as ions rushed across.
Internal voltage then would fall, and very soon be
lost.
But if conductances returned to resting states, anon,
A voltage would arise once more, predicted by Donnan.
This model was impressive as a formal working scheme.
And others came to increment a growing nervous team.
Lucas, Adrian, and Matthews. Gasser, Hartline, Forbes,
and Katz
Worked on nerves in many animals, from Limulns to
rats.
Spikes were soon a-popping in a host of different labs,
As the drive to learn their secrets grew intense and up
for grabs.
First, Hodgkin showed us how the nervous impulse moves
about:
Local circuits are the answer, sorting currents in and
out.
Then Cole and Curtis found that Bernstein's story had
some credence.
They showed that spikes in axons come from changes
in impedance.
And Hodgkin and his sidekick had the fun of being first
To learn potentials during spikes are actually reversed!
These two then went to Plymouth where, with Katz, they
made their camp;
And with their brains predicting gains, used KC's volt-
age clamp.
They needed nerves of super size to extricate their facts
from,
and J.Z. Young surprised them, saying "Squids have
giant axons."
These, threaded through with two fine twists of chloride-
coated wire
Were quickly held in voltage steps to stimulate their fire.
With capacitative surges voided early at each go.
The resulting current traces now gave signs of ion flow.
Now, sodium was first allowed sole access through its gate.
Then the membrane's charge would switch, a half a
millisecond late.
Potassium then rushed across, the other way about.
Until the membrane had reset its voltage in and out.
(Bernstein had it half-right, but he didn't know the sequel:
The channels all are separate — and their latencies aren't
equal)
Soon, Moore and Narahashi found a fishy substance
which,
When put on axons, blocked the voltage-gated sodium
switch.
167
168
D. MELLON
And TEA when tested in a voltage-clamped condition,
Removed the second ion flow: potassium emission.
Now, these two channels set the tune at every nervous
dance
But in the wings are many more, all waiting for their
chance.
Nerve terminals need calcium to talk about their states;
Membrane channels are the answer, controlled by volt-
age gates.
Other channels set the mood for the neurons' current
tempers.
Their gates are held in different states by second mes-
sengers.
And peptides that a year ago were totally unknown
Are recognized by schoolboys as the latest neurohor-
mone.
So now we know the ins and outs of axon current flow.
How membrane gates are modified to make the impulse
go.
The thoughts which built this story through the years,
with wire and prose.
Themselves were born of agencies they were destined
to disclose.
The dancing spikes thus spin their tale of learning, love,
and pain
And turning round, they watch it all come running
back again.
From: Earth. Robert H . and Robert E. Broshears. 1982. The Invertebrate World. Sounders College
Publishing. Philadelphia. P 301
Reference: Biol. Bull 182: 169-176. (April, 1992)
The Culture, Sexual and Asexual Reproduction, and
Growth of the Sea Anemone Nematostella vectensis
CADET HAND AND KEVIN R. UHLINGER
Bodega Marine Laboratory. P.O. Box 247, Bodega Bay, California 94923
Abstract. Nematostella vectensis, a widely distributed,
burrowing sea anemone, was raised through successive
sexual generations at room temperature in non-circulating
seawater. It has separate sexes and also reproduces asex-
ually by transverse fission. Cultures of animals were fed
Artemia sp. nauplii every second day. Every eight days
the culture water was changed, and the anemones were
fed pieces of Mytilus spp. tissue. This led to regular
spawning by both sexes at eight-day intervals. The cultures
remained reproductive throughout the year. Upon
spawning, adults release either eggs embedded in a gelat-
inous mucoid mass, or free-swimming sperm. In one ex-
periment, 12 female isolated clonemates and 12 male iso-
lated clonemates were maintained on the 8-day spawning
schedule for almost 8 months. Of the female spawnings,
75% occurred on the day following mussel feeding and
water change, and 64% of the male spawnings were sim-
ilarly synchronized under this regime. Fertilization and
development occur when gametes from both sexes are
combined in vitro. At 20°C, the embryos gastrulate within
12-15 hours. Spherical ciliated planulae emerge from egg
masses 36-48 hours post-fertilization. The planulae elon-
gate and form the first mesenteric couple, as well as four
tentacle buds, by day five. By day seven, they metamor-
phose and settle as 250-500 nm long, four-tentacled ju-
venile anemones. More tentacles and all eight macro-
cnemes are present at 2-3 weeks. Individuals may become
reproductively mature in as few as 69 days. Nematostella
vectensis has the potential to become an important model
for use in cnidarian developmental research.
Introduction
Many sea anemones can be maintained for long periods
under a variety of conditions including non-circulating
Received 25 July 1991; accepted 13 January 1992.
water at room temperatures (Stephenson, 1928), and un-
der the latter conditions some species produce numerous
asexual offspring by a variety of methods (Cary, 1911;
Stephenson, 1929). More recently this trait has been used
to produce clones of genetically identical individuals use-
ful for experimentation; i.e.. Haliplanella luciae(by Min-
asian and Mariscal, 1979), Aiptasia pulchella (by Muller-
Parker, 1984), and Aiptasia pallida (by Clayton and Las-
ker, 1984). We now add one more species to this list,
namely Nematostella vectensis Stephenson (1935), a small,
burrowing athenarian sea anemone synonymous with N.
pellucida Crowell (1946) (see Hand, 1957).
Nematostella vectensis is an estuarine, euryhaline
member of the family Edwardsiidae and has been recorded
in salinities of 8.96 to 5 1 .54%o and water temperatures of
- 1° to 28°C (Williams, 1983). It is a small animal, usually
less than 2 cm long and a few millimeters in diameter
when found in the field (Williams, 1983). It occurs in
England, from Nova Scotia to Georgia on the North
American Atlantic coast, from Florida to Louisiana along
the shores of the Gulf of Mexico, and from California to
Washington on the Pacific coast (Hand, unpub. Louisiana
record; Heard, 1982; Kneib, 1985; Williams, 1983). Wil-
liams (1983) considered the species vulnerable to extinc-
tion in Great Britain, but it is plentifully abundant
throughout most of its range and is readily collected. Ne-
matostella occurs in soft sediments, in plant debris, and
among living plants in permanent pools and tidal creeks
in salt marshes. It also occurs subtidally in estuaries in
Chesapeake Bay (M. Posey, pers. comm.; Calder, 1972)
and in the Indian River in Delaware (Jensen, 1974).
To date we have only the barest outline of the life history
of this species. Crowell (1946) and Frank and Bleakney
(1976) reported that eggs were discharged in mucoid
masses accompanied by numerous nematosomes. Ne-
matosomes, which occur in the coelenteron, are spherical,
15-45 jim, flagellated bodies containing nematocysts and
169
170
C. HAND AND K. R. UHLINGER
are known only from the genus Nematostella (Williams,
1979). Frank and Bleakney (1976) reported that planula
larvae developed from the eggs, but subsequently disap-
peared, and Williams (1975) found three, 1.0 mm long
planulae that he attributed to Nematostella in a pool con-
taining that sea anemone. Rudy and Rudy (1983) kept
N. vectensis in the laboratory for five years and stated that
eggs developed to planulae in three days and to "four-
knobbed" juveniles, i.e.. with four tentacle buds, in five
days. The sexes are separate (Hand, 1957; Frank and
Bleakney, 1976; Williams, 1975), and TV. vectensis repro-
duces asexually by transverse fission (Lindsay, 1975; Wil-
liams, 1976; Frank and Bleakney, 1978).
Even less is known about the natural history of N. vec-
tensis. Kneib (1985) has shown that the grass shrimp Pa-
laemonetes pugio may prey on this anemone, and Lindsay
(1975) and Frank and Bleakney (1978) have given us in-
formation on the anemone's diet. We also know that it
tolerates extremes of temperature and salinity (Bleakney
and Meyer, 1979; Stephenson, 1935) and, at times, may
occur in dense populations, i.e., over 5 million in a single
pool (Williams, 1983) and 1816 in a 15 cm2 sample
(Bleakney and Meyer, 1979). Little beyond this is known
of its natural history.
Here we describe the culture, reproduction, develop-
ment, and growth of Nematostella, as well as some other
aspects of its biology. In particular, we show that this ane-
mone reproduces sexually in standing water at room tem-
perature, is readily raised through successive generations,
is sexually active throughout the year, and shows no sign
of seasonality in its reproduction in the laboratory. This
combination of traits — namely asexual reproduction,
which allows the development of clones, and sexual re-
production with subsequent development through larval
stages to reproductive adults, all under room temperature
culture conditions — suggests that this sea anemone should
be useful in the study of cnidarian biology, particularly
development.
Materials and Methods
In December 1987, we received 12 living N. vectensis
that had been collected subtidally from the Rhode River,
a subestuary of the Chesapeake Bay in Maryland. The
largest of these anemones was about 1 5 mm long when
fully extended. The salinity at the time and place of col-
lection was about \2%o. These Rhode River anemones,
together with their sexual and asexual descendants, have
been maintained in our laboratory and now number sev-
eral thousand. It is from these cultures that isolated female
and male clonemates were reared (see below). We also
have cultures of Nematcstella from England, Nova Scotia,
Georgia, California, Oregon, and Washington.
Culture methods
Our cultures were maintained in crystallizing dishes
with plastic Petri dish parts as covers. They were kept at
room temperatures ranging from 16-26°C, and at a sa-
linity of about 12%o. We did not provide these animals
with any substrate, such as silt or fine sediments, nor did
we provide aeration to the cultures. The water was
changed weekly to bi-weekly, but solitary anemones or
cultures of only a few individuals may actually be kept
for several weeks in unchanged water. Nematostella will
tolerate crowding. We have raised about 300 sea anemones
to lengths of 2-4 cm in a single 80 X 40 mm dish con-
taining 100 ml of water, and we have reared equal num-
bers from planulae to young sea anemones, about 1 .0 cm
long, in 25 ml of water in 51 X 31 mm dishes.
We fed Artemia nauplii to our cultures every second
day, and cultures have been maintained for more than
two years on that diet alone; we have used both San Fran-
cisco Bay Brand and Sanders Premium Great Salt Lake
Anemia. Other foods used were the yolk of hard boiled
hens' eggs and veliger larvae of mussels and oysters. These
are readily accepted by recently metamorphosed sea ane-
mones. Tissues from Mytilus edulis and M. californiamis,
such as the ovary cut into 1-2 mm pieces, are also readily
eaten by larger Nematostella.
The production of isolated female and male clonemates
The 12 Nematostella received from the Rhode River
in December 1987 grew rapidly and began producing fer-
tile egg masses in February 1988. From these and sub-
sequent spawnings we reared several hundred Nemato-
stella to sexual maturity. To observe spawnings more
closely and to control the time of fertilization, we isolated
sibling anemones that were several months old. By April
1989 we had isolated 16 mature and reproductively active
females and 14 reproductive males. Each animal was held
in a 51 X 31 mm dish containing 25 ml of 33% seawater,
and each was fed 3-5 drops of concentrated Anemia nau-
plii every second day. Each animal was fed small pieces
of M. californiamis ovary every eighth day, the water in
each dish was changed irregularly, and the pattern of
spawning was observed.
In time, through asexual reproduction by transverse
fission, many of the isolated individuals became clonal
groups, and in the period from February 1989 to Decem-
ber 1989, one particular isolated male anemone became
a clone of 96 individuals and one female became a clone
of 38. From these two clones, we isolated 12 female and
12 male clonemates as above. These isolated anemones
were fed several drops of nauplii every second day and
two pieces of M. californiamis ovary every eighth day fol-
lowed by a water change. We recorded spawnings for these
anemones from 12 February to 3 October 1990 (Table I).
CULTURE OF NEMATOSTELL.4
171
Effects of salinity
Because Nematostella is euryhaline and because it re-
produced frequently for us, we explored the effect of sa-
linity on both sexual and asexual reproduction. We pre-
pared the following concentrations of seawater: 10%, 20%,
33%, 66%, 100%, 125%. The salinity of the 100% seawater
was 34%o, and the 125% seawater was prepared by evap-
oration. We selected 6 groups of 20 anemones each from
a culture of about 300, essentially mature, 6-month-old
siblings, 2.0-3.0 cm long. Other than the group of 20 that
was to stay in 33% seawater, each group was acclimated
to the desired final concentration by being successively
moved, every four days, through the increasing or de-
creasing concentrations. We fed these anemones brine
shrimp nauplii every second day, and recorded their sexual
and asexual reproduction for a sixteen week period, from
mid-October 1988, to the end of January 1989.
Results
Sexual reproduction
In our cultures, anemones become sexually mature at
three to four months of age and at column lengths of
between 1.5 and 3.5 cm. The sexes are separate, and in-
dividuals that have been isolated for more than two years
continue on as either males or females. We have seen no
signs of hermaphroditism or change of sex. In cultures of
mixed sexes, spawning frequently occurred in numerous
dishes on a given day; i.e., cultures on comparable feeding
regimes tended to spawn at the same time. Egg masses
formed within females are extruded through the mouth
(Fig. 1). The eggs are opaque and creamy white, and they
vary in diameter from 1 70 to 240 ^m. The masses consist
of a gelatinous-like material which adheres to nearby ob-
jects when first extruded. The masses may be small and
spherical (up to about 2 or 2.5 mm diameter) or elongate,
and in the extreme, more than 5 cm long by 3 mm in
diameter (Fig. 2). There may be few eggs, i.e., 5-10, such
as reported by Crowell ( 1 946), or there may be many more
(the largest egg masses we have seen contained more than
2000 ova). As well as ova, the egg masses contain
hundreds, even thousands, of nematosomes. These can
be seen rotating in place within the egg mass. In our cul-
tures, sexual reproduction has occurred in every month
of the year with no apparent seasonality or correlation
with moon phases.
In the first experiment with 16 female and 14 male
isolated siblings, we recorded numerous instances when
most of both sexes spawned within a few hours of one
another, between mid-afternoon and early evening. Fe-
males produced from one to three egg masses each
spawning, and males released varying amounts of sperm.
In the second experiment, with 1 2 female and 1 2 male
isolated clonemates, we recorded 322 female and 264 male
spawnings; 242 of those by females and 1 70 of those by
males occurred the day after both sexes had eaten mussel
and had had their water changed (Table I). Thus 75% of
the female spawnings and 64% of the male spawnings
occurred on the same days and, as before, within a few
hours of one another. Of those spawnings, all 1 2 females
spawned in seven cases, and all 1 2 males spawned in four.
On three occasions, all 1 2 of both sexes spawned on the
same day. On the day after eating mussel, at least one
female always spawned, but on three occasions, no male
spawned.
Embryology and development
Sperm produced by isolated males can be added to ex-
truded egg masses and development observed. Cleavage
leads to translucent blastulae, most of which become in-
vaginate gastrulae 12-15 hours after fertilization at around
20°C (Fig. 3). The gastrulae emerge from the egg mass as
200-250 nm spherical, ciliated planulae 36-48 hours after
fertilization. The planulae alternate between periods of
swimming and resting and develop an apical tuft of large
cilia that becomes obvious by the third day. They change
their shape progressively from spherical, to pear-shaped,
to elongate, and by five days, some develop four tentacle
buds around the mouth (Figs. 4, 5). At four to five days,
there are two thickened areas of tissue internally that rep-
resent the first mesenteric couple. By the seventh day,
many planulae cease swimming, settle to the bottom, and
metamorphose into 250-500 nm long juveniles with four
tentacles. The metamorphosed young may retain cilia on
their columns for more than a month and grow to a length
of more than 1 mm before the cilia are lost. During the
first few days after metamorphosis, the juveniles glide over
the substrate with the aboral end forward, although they
no longer rotate about their longitudinal axes as the plan-
ulae did. The direction of movement reverses after a few
days, and the juveniles then glide with the oral end leading.
Most juveniles cease gliding before they are 1 mm long.
The young anemones vary considerably in size, and by
10 days some may already be 1 mm long when fully ex-
tended. By two weeks some will have grown to 2 mm (Fig.
6), by three weeks to 4 mm or slightly longer, and, in the
extreme, to 2.5 cm long in a month. At 2-3 weeks, a
second set of four tentacles develops, and all eight ma-
crocnemes are obvious, although the first couple are much
larger than any of the others. This seeming dominance of
the first mesenteric couple is a feature that remains ob-
vious for the first several months. Commonly, month-old
animals have 1 2 tentacles, can extend their bodies to 1 -
2 cm, and possess a few nematosomes. Two-month-old
animals are approaching sexual maturity, are 2-5 cm long,
may have 16 tentacles, and have usually developed abun-
dant nematosomes. Some mature sexually and spawn at
172
C. HAND AND K. R. UHLINGER
CULTURE OF NEM.4TOSTELL.-i
173
an age of about 10 weeks. Spawning occurred in one cul-
ture that was only 69 days post fertilization. Asexual di-
vision by transverse fission also becomes common at about
10 weeks. The earliest fission noted was in a seven-week-
old individual that was almost 3 cm long. In about five
months, heavily fed animals can grow to expanded lengths
exceeding 16 cm, with physal diameters of 4-5 mm, and
tentacles 2-3 cm long.
In every group of developing sea anemones, we have
observed variations in timing and size of individuals. Not
all planulae metamorphose to juveniles in seven days,
and some delay metamorphosis for at least two weeks. In
one instance, planulae remained active for as long as 135
days, and in that time their size decreased such that the
last one measured, just six days before it was last seen,
was about 100 nm long. Frequently a few planulae, 1%
or less, remain active for 1-2 months in bowls with their
developing siblings, but we do not know whether these
are still capable of metamorphosing.
As well as variations in growth rates, we have observed
newly metamorphosed juveniles with two and three ten-
tacles rather than the normal four. At ages of several
months to a year or more, there may be large variations
in the abundance of nematosomes. Too, some individuals
have large physal regions or very long tentacles compared
to others, and the frequency of asexual reproduction varies
greatly from individual to individual. There also may be
much variability in planular size, because the planulae in
our cultures seldom exceed 500 ^m long, a size substan-
tially less than those reported by Frank and Bleakney
(1976) and Williams (1975).
Nematosomes
Nematosomes are equally abundant in both sexes.
Those embedded in the egg masses emerge from the ma-
trix along with the emerging planulae. They do not move
throughout the water column, but tend to remain rotating
near the degenerating matrix of the original egg mass.
However, both the egg mass matrix and the nematosomes
may remain in the dish with the developing anemones
for extended periods. We have had nematosomes remain
active for as long as 1 3 days past the date of spawning,
and the gelatinous matrix from the egg mass, although
shrinking in size, may remain for a month or more.
Other populations
We have kept cultures from areas other than Chesa-
peake Bay on feeding and water changing schedules iden-
tical to those from Chesapeake Bay. These cultures also
tend to spawn synchronously with those from Chesapeake
Bay. The development, metamorphosis, and growth of
the offspring of those cultures do not differ from those of
the Chesapeake Bay anemones.
Salinity
The anemones in 10% and 20% seawater did not do
well; we terminated these two cultures at 5 weeks because
18 of the 20 in 10% seawater, and 13 of the 20 in 20%
seawater, were deflated and had mesenteries everted
through their mouths. There had been one asexual divi-
sion in the group in 20% seawater. The anemones in the
other salinities all produced fertile egg masses and planula
larvae, and all planulae, except those in 125% seawater,
metamorphosed to young anemones. At the end of 16
weeks (Table II), we discontinued this study. The ane-
mones in 33% seawater had grown to be 4-6 cm long,
had spawned four times, and by asexual reproduction had
become a group of 29 anemones. The group in 66% sea-
water did not grow much, and were barely larger than at
the initiation of the experiment. These had spawned four
times and had become a group of 28 animals. The ane-
mones in 1 00% seawater had decreased in size, the largest
being about 2.5 cm long when fully extended. These ane-
mones had become a group of 26 and had spawned only
once. The anemones in 125% seawater also only spawned
once, had become a group of 22, and decreased in size,
the largest being about 2.0 cm long.
Discussion
Nematostella vectensis is regarded as a small sea ane-
mone, and Williams (1983) stated that, although they may
Figure 1. Spawning female releasing part of an egg mass. Note the remaining unreleased egg mass in
colenteron at the arrow. Scale bar: 1 .0 cm.
Figure 2. Egg masses from numerous individuals collected from one evening's spawn. Scale bar:
1.0cm.
Figure 3. Blastulae. early gastrulae and nematosomes in situ in an egg mass fertilized 14 hours earlier.
Arrow points to a nematosome. Scale bar: 100 ^m.
Figure 4. Three day post-fertilization planula with early apical tuft. Scale bar: 100 fjm.
Figure 5. Five day post-fertilization planula with fully developed apical tuft and developing tentacle
buds. Tentacle bud at arrow. Scale bar: 100 ^m.
Figure 6. Two-week-old, four tentacled juvenile anemone. Arrow points to one member of the first
couple of mesenteries. Note its size compared to the adjacent smaller primary mesenteries. Scale bar:
1.0 mm.
174
C. HAND AND K.. R. UHLINGER
Table I
Spawning of 12 isolated female and 12 isolated male donemates of Nematostella vectensis from 12 February to 3 October 1990
Females
Number of spawns of anemone
Total
Day of spawn
1
2
3
4
5
6
7
8
9
10
11
12
spawns
%
Day before eating mussel
3
4
2
3
1
2
4
2
3
3
.1
4
32
10
Day of eating mussel
1
0
0
0
1
0
0
0
0
0
1
1
4
1
Day after eating mussel
20
18
20
21
21
22
19
21
21
19
19
21
242
75
Other days
3
3
5
4
3
5
3
2
7
3
2
4
44
14
Sum
27
25
27
28
26
29
26
25
31
25
23
30
322
100
Males
Number of spawns of anemone
Total
Day of spawn
1
2
3
4
5
6
7
8
9
10
11
12
spawns
%
Day before eating mussel
0
0
1
0
0
0
0
1
0
0
0
0
2
1
Day of eating mussel
6
2
6
6
4
5
6
3
7
5
8
3
61
23
Day after eating mussel
16
13
15
12
16
15
11
13
1 1
17
16
15
170
64
Other days
1
5
2
1
0
2
2
6
5
1
4
2
31
12
Sum
23
20
24
19
20
22
19
23
23
23
28
20
264
100
All anemones were fed brine shrimp nauplii every second day. Each eighth day they were fed pieces of Mytilus californianus ovary and the water
was changed. All were mature, 6-10 cm long adults at the initiation of the test and were maintained in 33% seawater.
be up to 6 cm long, they are usually less than 2 cm. We
were surprised, therefore, when our laboratory specimens
grew to more than twice the maximum size reported pre-
viously. All earlier size measurements appear to have been
made on recently collected animals, and not well-fed cul-
tured ones. The small size of the sea anemones in the
field, relative to the larger sizes in our cultures, must reflect
the small amount of food they capture in their native
habitats.
The production of gelatinous egg masses by Nemato-
stella is a unique feature of this sea anemone, although
the eggs of Halcampa ditodecimcirrata, which are released
individually, become surrounded by a jelly envelope after
fertilization (Nyholm, 1949). The jelly attaches the eggs
to the sandy bottom in which Halcampa lives.
We know of no other sea anemone that spawns re-
peatedly overextended periods, although an annual period
of reproductive activity is known for many sea anemones
(Jennison, 1979). N. vectensis may well have an annual
reproductive cycle in nature; but in the laboratory it has
spawned repeatedly and on a predictable schedule. We
have tried feeding mussel tissue every fourth day to some
female clonemates of the anemones on the eight-day cycle.
The results led to spawns in an erratic and unpredictable
fashion; apparently N. vectensis cannot spawn repeatedly
at four-day intervals. We now are attempting a seven-day
cycle, and early results suggest that predictable spawnings
will occur at seven-day intervals.
How the reproductive cycle of N. vectensis operates in
nature is unknown. Most populations of this anemone
Table II
Asexual reproduction, growth, and spawning during 16 weeks in various concentrations of seawater
Sea water
Initial
Final
Largest
Number of
Larvae
concentration
number
number
final size
spawns
metamorphosed
33%
20
29
6.0cm
4
yes
66%
20
28
3.5 cm
4
yes
100%
20
26
2.5cm
1
yes
125%
20
22
2.0cm
1
no
All anemones were 2.0-3.0 cm long, six-month-old siblings at the initiation of the experiment. All were fed brine shrimp nauplii every second
day, and the water was changed every one to two weeks.
CULTURE OF NEMATOSTELL4
175
live in pools in marshes at tidal elevations that do not
necessarily receive fresh water with each tidal cycle, and
their food consists of denizens of the pools they inhabit
(Lindsay, 1975; Williams, 1976; Frank and Bleakney,
1978). The higher tides generally provide fresh seawater
to the pools, and at times that water must carry large
amounts of plankton. We wonder whether the pulses of
extra food, in the form of the mussel ovary that we supply,
may mimic pulses of extra food from the plankton that
they receive in nature. Perhaps that pulse of food, along
with the change of water, is the key to the release of ga-
metes in N. vectensis.
The planula larvae of TV. vectensis, from the age of about
three days onward, are active swimmers, although they
do spend long periods immobile on the bottoms of our
culture dishes. Some swimming is spontaneous, but if the
cultures are disturbed, most of the motionless planulae
leave the bottom and swim actively. They swim in a
clockwise spiral, as viewed from the oral end of the plan-
ula, and while doing this they rotate around their longi-
tudinal axes in a clockwise direction. Widersten (1968)
reported similar rotation in several cnidarian larvae, in-
cluding several species of sea anemones, although he also
observed some anemone planulae that rotated either
clockwise or counterclockwise. In contrast to the generally
clockwise rotation of the sea anemone larvae, he found
only counterclockwise rotation in hydrozoan and scy-
phozoan planulae.
The reversal in the direction of gliding by the juvenile
anemones was unexpected and is previously unreported.
We presume that the same cilia that move the planulae
in an aboral direction later reverse their beat and move
the recently metamorphosed anemones in an oral direc-
tion. But the juveniles could alternatively be propelled by
newly developed cilia.
In the only specific study of nematosomes, Williams
(1979) found no correlation between the size of an ane-
mone and the number of its nematosomes, and we agree.
He also considered nematosomes to be functionless, and
although we find that difficult to accept, their function is
certainly not obvious. In his study, Williams also showed
that nematosomes removed from anemones had relatively
short lives; i.e., only those maintained at low temperature
(1.5-3.5°C) lived as long as 55 hours. In sharp contrast,
we found that nematosomes would live for 1 3 days outside
of the body of anemones at temperatures around 20°C.
Williams (1979) made his observations on material in
normal seawater (34%o), whereas our material was in a
salinity of about 12%o. Too, we did not artificially free the
nematosomes from the anemones. Our observations were
on nematosomes contained in egg masses, from which
they emerged along with the planulae.
As we noted earlier, N. vectensis is a euryhaline sea
anemone; it can be found in widely varying salinities. Our
own experience in culturing this anemone confirms that
we are dealing with a widely tolerant euryhaline, eury-
thermal sea anemone, and we know of no other sea ane-
mone of equal tolerance. Not only is N. vectensis tolerant,
but it carries out its full repertoire of sexual and asexual
reproduction, development, and growth in a wide range
of salinities. When cultured in full strength seawater, or
higher salinities, growth seems to be inhibited, and we
have not observed successful sexual reproduction and
subsequent metamorphoses in salinities greater than 34%».
Sexual reproduction and the subsequent development
in sea anemones have rarely been studied (see review by
Stephenson, 1928; also Mergner, 1971; Campbell, 1974;
and Fautin et ai. 1989). Those species that have been
reported upon had all been recently collected, brought in
to a nearby laboratory or field station, and subsequently
spawned. Fortuitously, investigators on the scene, such
as Nyholm (1943, 1949), Chia and Spaulding (1972), Sie-
bert (1973, 1974). Riggs (1988), and Chia et al. (1989),
have been able to examine some of the events from
spawning onwards, although until now there have been
no reports of rearing of successive generations of any spe-
cies. Most studies do not describe development beyond
the planula, although those of Spaulding (1972) on
Peachia quinquecapitata. Chia and Spaulding (1972) on
Tealia crassicornis, and Siebert (1973) on Stomphia di-
demon are exceptions. Of all the studies of development
of sea anemones, we are aware of only two (Clark and
Dewell, 1974; Larkman and Carter, 1984) that have pro-
vided a close look at fertilization and related events. Those
studies, like the others we have cited on matters related
to development, were largely made possible by sponta-
neous spawnings and not as the result of planned or con-
trolled spawnings. Now that we can culture Nematostella
throughout its life history and can control some of the
variation through the use of clonal anemones, we can
look closely and repeatedly at all events, from fertilization
to spawning. Moreover, the short generation time of two
to three months in laboratory-reared Nematostella will
allow ready genetic analyses. Because Nematostella can
be cultured away from marine facilities, research on every
aspect of its life history can be carried out at inland lab-
oratories, and we believe this sea anemone has the poten-
tial to become an important model for research in cni-
darian biology.
Acknowledgments
We are grateful to Martin Posey for the material from
Chesapeake Bay, and we also owe our thanks to the fol-
lowing for other material: Martin Sheader and Daphne
Fautin for specimens from England; Sherman Bleakney
for specimens from Nova Scotia; R. T. Kneib for speci-
mens from Georgia; Edward Lyke, Daniel Wickham, and
176
C. HAND AND K. R. UHLINGER
Pamela Roe for specimens from California; Jon Geller
for specimens from Oregon; Eugene Kozloff, Edward
Lyke, and Claudia Mills for specimens from Washington.
We thank Tzyy-ing Chen, Jennifer Russo, and Eleanor
Uhlinger for their help in caring for our animals, and we
are grateful for the assistance Beth Clark provided in the
preparation of our manuscript. Fred Griffin and Eduardo
Almeida each provided invaluable assistance with pho-
tography and we thank Wallis Clark for the use of his
laboratory's photographic and optical equipment. Com-
ments and suggestions from Wallis Clark, Fred Griffin,
and Eleanor Uhlinger have been of great assistance in the
development of the manuscript. This work is a result of
research sponsored in pan by NOAA, National Sea Grant
College Program, Department of Commerce, under grant
number NA89AA-D-SG138, project number 83-A-N,
through the California Sea Grant College. The U.S. Gov-
ernment is authorized to reproduce and distribute reprints
for governmental purposes.
Literature Cited
Bleakney, J. S., and K. B. Meyer. 1979. Observations on saltmarsh
pools, Minas Basin, Nova Scotia 1965-1977. Proc. N.S. Inst. Sa 29:
353-371.
Calder, D. 1972. Phylum Cnidaria. Pp. 97-107 in A Check List of the
Biota of Lower Chesapeake Bay Special Scientific Report No. 65,
Virginia Institute of Marine Science.
Campbell, R. D. 1974. Cnidana. Pp. 133-199 in Reproduction of Ma-
rine Invertebrates, I'ol. 1. Acoelomate and Pseudocoe/omate Meta-
zoans, A. C. Giese and J. S. Pearse. eds. Academic Press, New York.
Cary, L. R. 1911. A study of pedal laceration in actinians. Bio/. Bull
20:81-108.
Chia, F.-S., and J. G. Spaulding. 1972. Development and juvenile
growth of the sea anemone Tealia crassicornis. Biol. Bull. 142: 206-
218.
Chia, F.-S., J. Liitzen, and I. Svane. 1989. Sexual reproduction and
larval morphology of the primitive anthozoan Gonaclinia prolifera
M. Sars. J. Exp. Mar. Biol Ecol. 127: 13-24.
Clark, W. H., Jr., and W. C. Dewel. 1974. The structure of the gonads.
gametogenesis, and sperm-egg interactions in the Anthozoa. Am. Zool.
14: 495-510.
Clayton, W. S., Jr., and H. R. Lasker. 1984. Host feeding regime and
zooxanthellal photosynthesis in the anemone, Aiptasia pallida ( Ver-
rill). Biol. Bull. 167: 590-600.
Crowell, S. 1946. A new sea anemone from Woods Hole. Massachu-
setts. J Wash. Acad. Sci. 36: 57-60.
Fautin, D. G., J. G. Spaulding, and F.-S. Chia. 1989. Cnidaria. Pp.
43-62 in Reproductive Biology of Invertebrates, I'ol. II'. Pan A.
K. A. Adiyodi and R. G. Adiyodi, eds. John Wiley & Sons, Chichester.
Frank, P., and J. S. Bleakney. 1976. Histology and sexual reproduction
of the anemone \ematostella vectensis Stephenson 1935. J. Nat. Hist.
10: 441-449.
Frank, P., and J. S. Bleakney. 1978. Asexual reproduction, diet and
anomolies of the anemone Nematoslella vectensis in Nova Scotia.
Can. Field-Nat. 92: 259-263.
Hand, C. 1957. Another sea anemone from California and the types
of certain California species. J. ll'axlt. Acad. Sci. 47: 41 1-414.
Heard, R. W. 1982. Guide to common tidal marsh invertebrates of
the northern Gulf of Mexico. Mississippi Alabama Sea Grant Con-
sortium. Publ. 79-004. 82 pp.
Jennison, B. L. 1979. Gametogenesis and reproduction cycles in the
sea anemone Anthopleura eleganlissima (Brandt, 1835). Can. J. Zool.
57:403-411.
Jensen, L. D. 1974. Environmental responses to thermal discharges
from the Indian River Station. Indian River, Delaware. Electric Power
Research Institute. Report No. 12. Publ. No. 74-049-00-3.
Kneib, R. T. 1985. Predation and disturbance by grass shrimp. Palae-
moneles pugio Holthuis, in soft-substratum benthic invertebrate as-
semblages. J. Exp. Mar. Biol. Ecol. 93: 91-102.
Larkman, A. LI., and M. A. Carter. 1984. The apparent absence of a
cortical reaction after fertilization in a sea anemone. Tissue & Cell
16: 125-130.
Lindsay, J. A. 1975. A salt marsh anemone. Mar Aquarist 6(8):
43-48.
Mergner, H. 1971. Cnidaria. Pp. 1-84 in Experimental Embryology
of Marine and Freshwater Invertebrates. G. Reverberi. ed. North
Holland Press, Amsterdam.
Minasian, L. L., Jr., and R. N. Mariscal 1979. Characteristics and
regulation of fission activity in clonal cultures of the cosmopolitan
sea anemone. Haliplanella luciae ( Verrill). Biol. Bull. 157: 478-493.
Muller-Parker, G. 1984. Photosynthetic-irradiance responses and
photosynthetic periodicity in the sea anemone Aiptasia pulchella and
its zooxanthellae. Mar. Biol. 82: 225-232.
Nyholm, K. G. 1943. Zur Entwicklung und Entwicklungsbiologie der
Ceriantharien und Aktinien. Zool. Bidr. Uppsala 22: 87-248.
Nyholm, K. G. 1949. On the development and dispersal of Athenaria
actinia with special reference to Halcampa duodecimcirrata M. Sars.
Zool. Bidr L'ppsala 27: 465-506.
Riggs, L. L. 1988. Feeding behavior in Aiptasia tagetes (Duchassaing
and Michelotti) planulae: a plausible mechanism for zooxanthellae
infection of aposymbiotic planktotrophic planulae. Caribb. J. Sci.
24:201-206.
Rudy, P., Jr., and L. H. Rudy. 1983. Oregon Estuarine Invertebrates:
A n Illustrated Guide to the Common and Important Invertebrate An-
imals. Fish and Wildlife Service. FWS/OBS-83/16.
Siebert, A. E., Jr. 1973. A description of the sea anemone Stomphia
dtdemon sp. nov. and its development. Pac. Sci. 27: 363-376.
Siebert, A. E., Jr. 1974. A description of the embryology, larval de-
velopment, and feeding of the sea anemones Anlhopleura elegantis-
simaandA. xanihogrammica. Can. J. Zool. 52: 1383-1388.
Spaulding, J. G. 1972. The life cycle of Peach ia quinqiiecapitata, an
anemone parasitic on medusae during its larval development. Biol.
Bull. 143: 440-453.
Stephenson, T. A. 1928. The British Sea Anemones. I'ol. I The Ray
Society. London.
Stephenson, T. A. 1929. On methods of reproduction as specific char-
acters. / Mar. Biol Assoc. I'.K. 16: 131-172.
Stephenson, T. A. 1935. The British Sea Anemones. I W. // The Ray
Society, London.
Widersten, B. 1968. On the morphology and development in some
cnidarian larvae. Zool. Bidr. i'ppsala 37: 139-182.
Williams, R. B. 1975. A redescription of the brackish-water sea ane-
mone Nematoslella vectensis Stephenson, with an appraisal of con-
generic species. J. Nat. Hist 9: 51-64.
Williams, R. B. 1976. Conservation of the sea anemone Nematostella
vectensis in Norfolk, England and its world distribution. Trans. Nor-
folk Norwich Nat. Soc. 23: 257-266.
Williams, R. B. 1979. Studies on the nematosomes of Nematostella
vectensis Stephenson (Coelenterata, Actiniaria). J. Nat. Hist. 13: 69-
80.
Williams, R. B. 1983. Starlet sea anemone: Nematostella vectensis.
Pp. 43-46 in The IUCN Invertebrate Red Data Book. IUCN, Gland,
Switzerland.
Reference: Biol. Bull. 182: 177-187. (April. 1992)
Morphology and Development of a Unique Type of
Pelagic Larva in the Starfish Ptemster tesselatus
(Echinodermata: Asteroidea)
LARRY R. MCEDWARD
Department of Zoology, University of Florida, Gainesville, Florida 3 26 11
Abstract. Several unusual features characterize the
morphology of the pelagic larva of the starfish Pteraster
tesselatus and its metamorphosis into the juvenile stage:
( 1 ) morphogenesis of the supradorsal membrane during
metamorphosis by fusion of 1 5 lobes on the aboral region
of the body; (2) absence of brachiolar arms and attachment
disk; (3) heterochronic acceleration of development in the
water vascular system, and use of podia for attachment
to the substratum at settlement; (4) radial (rather than
bilateral) symmetry of the larva; and (5) congruent larval
and adult axes of symmetry, and a transverse orientation
of the adult rudiment within the larva. Collectively, these
features demonstrate that P. tesselatus has a highly derived
mode of development and a larva that is unique among
the asteroid echinoderms. In contrast to the current in-
terpretation of this larva as a modified pelagic brachiolaria,
I suggest that the unusual larva of Pteraster represents an
example of an apparently rare evolutionary transition in
animal development: the re-evolution of pelagic larval
development from benthic brooding.
Introduction
Major evolutionary transitions in the patterns of animal
development are rare. Some transitions, such as the
change from feeding larval development to nonfeeding
larval development, involve such drastic morphological
reorganization that the transition is irreversible (Strath-
mann, 1978). The evolution of pelagic larval development
from benthic brooding has not been documented in the
echinoderms. This suggests that, once lost from a clade,
a pelagic dispersive stage might not be re-evolved. This is
an important issue because it bears on the evolutionary
Received 6 November 1991; accepted 22 January 1992.
flexibility of the marine fauna and the predicted conse-
quences of major environmental change: faunal turnover
due to extinction and replacement or adaptive response
through modification of development and life history. In
this paper I describe the morphology and development of
a unique type of pelagic larva from the starfish Pteraster
tesselatus. The unusual features of this larva are consistent
with the hypothesis that it has evolved from benthic
brooding, rather than by modification of a pelagic larva
(Strathmann, 1974). The significance of this finding is
that Pteraster could provide a system for elucidating the
functional and developmental changes that underlie the
transition from benthic to pelagic development.
Starfish in the family Pterasteridae possess an unusual
structure known as the supradorsal membrane. In most
pterasterids, the supradorsal membrane is a thick (1-2
mm) layer that extends aborally from the lateral margins
of the ambulacra to form a secondary covering over the
body wall of the arms and disk (Fig. 1). This structure is
supported above the body wall by skeletal elements (pax-
illae) and encloses a space, the nidamental chamber. In
Pteraster tesselatus, the supradorsal membrane lacks
skeletal ossicles, but contains muscles and mucus cells, is
perforated by numerous minute spiracles, and possesses
a single large osculum (Fig. 1 ) located in the center of the
aboral surface( I ves, 1 888; Fisher, 191 1, pp. 355-363; Ro-
denhouse and Guberlet, 1946; Verrill, 1914, pp. 268-269).
The supradorsal membrane has three known functions:
respiratory ventilation, defense, and reproduction. Alter-
nating muscular contractions of the supradorsal mem-
brane and the aboral body wall produce rhythmic expan-
sions and contractions of the nidamental chamber. Sea-
water enters the chamber via ambulacra! pores, flows over
the respiratory papulae, and exits via the osculum (Jo-
hansen and Petersen, 1971; Nance and Braithwaite, 1981).
177
178
L. R. McEDWARD
Figure 1. Adult specimen of Pleraster tesselatus Ives 1888. Aboral
view. Scale bar = 2 cm. Supradorsal membrane (SDM) removed from
two arms and part of the disk, exposing the aboral body wall (BW).
Osculum (O) located center of aboral surface of supradorsal membrane.
In addition, the supradorsal membrane contains numer-
ous mucous cells associated with the spiracles (Roden-
house and Guberlet, 1946). Production of tremendous
quantities of mucus occurs in association with expulsion
of water from the nidamental chamber through the spi-
racles rather than the osculum. The mucus effectively de-
ters predation by the starfish So/aster dawsoni and Pyc-
nopodia helianthoides, possibly because of the presence
of saponin-like compounds (Nance and Braithwaite,
1979). Finally, the nidamental chamber is used for brood
protection (Keren and Danielssen, 1856; McClary and
Mladenov, 1988). In many species of pterasterids, young
are retained within the nidamental chamber of the mother
throughout development to the juvenile stage (e.g., Pier-
aster obscurus, Fisher, 1911, p. 363-368; Verrill, 1914,
pp. 274-277; Pteraster stellifer. Fisher, 1940, pp. 199-
200; Diplopteraster verrucosus. Fisher, 1940, pp. 201-203;
Hymenaster praecoquis, Sladen, 1889, pp. 524-525.
Benthic, brooding development has been considered
the rule in pterasterids. Fisher (1940, p. 73) stated that
probably all species in the genus Pteraster were brooders.
However, Chia (1966) reported that P. tesselatus spawned
eggs, and he provided a brief description of their devel-
opment as pelagic larvae. In addition, Pteraster militaris,
which broods embryos (Kaufmann, 1968; McClary and
Mladenov, 1989, 1990), also spawns some eggs that pre-
sumably develop as pelagic larvae (McClary and Mlad-
enov, 1988). But, beyond the observation that pelagic de-
velopment occurs within the genus, very little is known
about the morphology and development of Pteraster
larvae.
Here I report the results of a re-examination of the
development of Pteraster tesselatus in which I either dis-
covered or reinterpreted a number of unusual features:
( 1 ) morphogenesis of the supradorsal membrane during
metamorphosis, (2) absence of brachiolar arms and at-
tachment disk, (3) heterochronic acceleration of devel-
opment in the water vascular system, (4) radial symmetry
of the larva, and (5) congruent larval and adult axes of
symmetry. Because of these features, the larva of Pteraster
tesselatus is unlike that reported in any other asteroid.
This description of the morphology of the pelagic larva
and its development into the benthic juvenile will provide
the basis for investigating the evolution of this unusual
pattern of larval development (McEdward, in prep.).
Materials and Methods
Pteraster tesselatus Ives 1888 (Order Velatida; see
Blake, 1987) is a subtidal, often deep-water starfish that
occurs along the Pacific coast of North America from
central California to the Bering Sea (Lambert, 1981, p.
88). SCUBA was used in the collection of adult starfish
from depths of 5-20 m at several sites near the Bamneld
Marine Station (48°49TM, 125°08'W) in Barkley Sound,
Vancouver Island, British Columbia, Canada, and from
depths of 1 5-30 m near the Friday Harbor Laboratories
(48°32'N, 123°0'W) in the San Juan archipelago, Wash-
ington. Pteraster tesselatus is reproductive during July
and August (Chia, 1966; McEdward, pers. obs.).
Females were induced to spawn by intracoelomic in-
jection of 2-5 ml (10~4 A/) of the hormone 1 -methyl ad-
enine. Eggs were released within 1-3 h after injection and
developed without artificial insemination.
Embryos, larvae, and juveniles were cultured in plastic
beakers equipped with mesh bottoms (500 ^m mesh). The
beakers were suspended from a rack in an aquarium with
flowing seawater (see descriptions in Hoeg, 1984; Strath-
mann, 1987, p. 15). The mesh bottom allowed continuous
exchange of seawater between the aquarium and the cul-
ture containers. The seawater was not filtered. After the
larvae hatched, =250-300 healthy larvae were pipetted
from each culture into clean mesh-bottom beakers for
subsequent rearing.
Light microscope photographs and most of the original
observations were made on living embryos, larvae, and
juveniles. In some cases, larvae and juveniles were fixed
and cleared to render them transparent for observation
of internal features, such as the water vascular system and
the skeleton. Specimens were fixed in 10% formalin in
seawater (10 min), then dehydrated stepwise in ethanol
(30%, 50%, 70% twice, 90%, 100%, 2 min each). Contrast
was enhanced by staining the specimens with borax car-
A UNIQUE TYPE OF STARFISH LARVA
179
Figure 2. Light micrographs of egg and embryos of P/eraster lesse-
latus. Scale bars = 0.2 mm. (Top) Newly spawned egg. Jelly coat (JC)
surrounds the egg. The vitelline layer lies between the jelly and the egg
mine (2 min) between the two changes of 70% ethanol.
Immediately following dehydration, the specimens were
transferred to one of three clearing agents: methyl salic-
ylate (= oil of wintergreen), clove oil, or a mixture of
benzyl alcohol and benzyl benzoate (range 1:3-3:1).
Clearing was complete within 30 min to 1 h. Specimens
were observed and stored in the clearing agent.
Specimens were fixed for scanning electron microscopy
in cold 2% osmium tetroxide (1 h) in 0.45 ^m filtered
seawater, rinsed twice with distilled water, dehydrated
through a graded series (30%, 50%, 70%, 15 min each) of
ethanol, and stored in 70% ethanol. In preparation for
drying, specimens were dehydrated stepwise to absolute
ethanol (90%, 100%, 15 min each), then infiltrated with
hexamethydisilazane (HMDS, Sigma Chemical Co.) for
several hours. Specimens were air-dried at room temper-
ature (Nation, 1 983) in a dust-free chamber, sputter coated
with gold-palladium, and stored under desiccation.
Results
Pelagic larval development
Eggs were spawned from interradial gonopores into the
nidamental chamber and carried out through the osculum
of the supradorsal membrane with the exhalant flow of
water. Usually between 1-10 eggs were released with each
ventilation. The eggs were large, ranging in size from 1 .0
to 1 .4 mm in diameter (Fig. 2 A). They were opaque, yolky.
and positively buoyant. Egg color varied among spawns
from light yellow to dark red, but was most commonly a
rich orange color. There were not obvious correlations
among egg color, egg size, or the success of development.
The eggs were surrounded by a thick (> 100 yum) jelly coat
(Fig. 2 A). The jelly coat was lost prior to hatching, typically
within the first 48 h of development (see Table I for chro-
nology of development).
The cleavage pattern was variable and irregular, not
the typical radial pattern characteristic of asteroids.
Cleavage led to the formation of a blastula that initially
had a smooth wall. With continued division of the blas-
tomeres, the blastular wall was deeply folded to produce
a wrinkled blastula (Fig. 2B). Gastrulation occurred within
the vitelline envelope and involved the formation of a
broad, shallow archenteron with a large blastopore at the
vegetal pole (Fig. 2C). Archenteron formation was cor-
related with the loss of folding of the blastular wall in the
vegetal hemisphere of the embryo. Subsequently, the
cell membrane but is not visible until after loss of the jelly coat. (Middle)
Wrinkled blastula. Age = 36 h. Blastular wall with deep fold (F). Note
that this is a preserved specimen and the jelly coat and vitelline layer
have swollen. (Bottom) Gastrula. Vegetal view. Age = 2 days. Large
blastopore (B) lies at the center of the vegetal pole.
180
L. R. McEDWARD
Table I
Chronology of development in the starfish Pteraster tesselatus
at 1I-13°C
Age (days)
Developmental stage or event
0 Spawning; initiation of development
2 Wrinkled blastula; gastrulation
3 Hatching; ovoid, ciliated, swimming larva
5 Circumferential groove divides larva into two body
regions
7 Podia visible within circumferential groove
8 Five marginal bulges form in posterior body region
10 Settlement; ten marginal lobes present, anterior yolky
region flattened
1 1 Five aboral lobes begin to form
1 3 Aboral lobes well developed
1 7 Fusion of 1 5 lobes begins
19 Fusion of lobes complete
20 Some individuals still swimming
28 Supradorsal membrane smooth, no indication of lobes
2 (mo.) Mouth functional
2.5 (mo.) Distinct arms extend beyond margin of disk
blastopore was greatly reduced in size, and this occurred
in association with a progressive loss of folding of the
blastular wall, from the equatorial region of the vegetal
hemisphere towards the animal pole. The loss of folding
was completed by the time hatching from the vitelline
envelope occurred.
Prior to hatching, the eggs and embryos floated at the
surface. After hatching, larvae swam actively through the
water column. Anterior was denned as the end that was
directed forward during swimming. Hatching yielded a
simple, ciliated, ovoid larva (1.1-1.2 mm length). Distinct
ciliated bands for food capture were absent, and the cilia
remained uniformly distributed over the surface of the
larva throughout development (Fig. 3 A). A larval mouth
and functional gut were absent, and the blastopore closed
soon after hatching; development was entirely lecitho-
trophic.
Within 1-2 days of hatching, a circumferential groove
(Chia, 1966) formed around the larva, sa'A-'/s of the way
back from the anterior end (Fig. 3A). The groove divided
the larva into two distinct body regions: an anterior region
that contained mostly nutritional stores and was resorbed
during development, and a posterior region that developed
into the juvenile starfish. In striking contrast to most other
asteroid larvae, specialized settlement structures (bra-
chiolar arms and attachment disk; see Fig. 7) did not form
in the anterior region at any time during development
(Figs. 3, 5A).
Shortly after the formation of the circumferential
groove, the posterior region of the larva shortened along
the anterior-posterior axis. At the same time, five broad
marginal bulges formed around the circumference of the
larva, immediately posterior to the groove (Fig. 3B). The
larva continued to shorten over the next 1-2 days until
it was =0.9-1.0 mm in length. The posterior region as-
sumed a domed shape that was pentagonal in outline be-
cause of the broad marginal bulges. The region of the
larva anterior to the groove assumed a flattened, plate-
like shape (Figs. 3B, 5A).
At about the same time that the marginal bulges first
became visible, podia emerged from the groove (Fig. 3B).
The podia were distributed around the circumference of
the larva in five clusters, each located close to the center
of one of the marginal bulges. Within one week of hatch-
ing, there was an unpaired terminal podium and two pair
of functional podia in each cluster. The first pair of podia
Figure 3. Scanning electron micrographs of pelagic larvae of Pteraster
tesselatus. Scale bars = 0.2 mm. (Top) Lateral view. Age = 6 days.
Circumferential groove (CG) divides the larval body into anterior (AR)
and posterior (PR) regions. Note the remnant of the closing blastopore
( B) at the posterior end. (Bottom) Lateral view. Age = 8 days. The terminal
unpaired podium and the first pair of podia (P) of two ambulacra are
visible within the circumferential groove. The bilobed marginal bulges
(MB) can be seen just posterior to the groove.
A UNIQUE TYPE OF STARFISH LARVA
181
was considerably longer than the second pair. The clusters
became less distinct as these podia developed and addi-
tional podia formed. Eventually the podia came to be
distributed in a ring around the circumference of the larva,
within the groove (Fig. 5C).
Settlement and metamorphosis
Because Pteraster lacks purely larval structures char-
acteristic of asteroids (e.g., ciliated bands, gut, or bra-
chiolar arms) and does not undergo a well-defined period
of metamorphosis, larval and juveniles stages cannot be
rigorously defined using morphological or developmental
criteria. Therefore, I have used ecological criteria: "larva"
refers to the free-swimming, pelagic, dispersive stage of
the life cycle and "juvenile" refers to the animal following
initial settlement and assumption of the adult orientation
on the substratum. In fact, the fully formed juvenile star-
fish was not achieved until weeks to months after settle-
ment.
Initially larvae swam with the anterior end forward.
Later, as the rudiment of the juvenile starfish developed
in the posterior end, the orientation of the larval body in
the water column changed and the yolky anterior end was
directed upward. During settlement, the larva attached to
the substratum using podia. However, attachment could
only occur when the larva turned on edge because the
podia were not long enough to reach around the posterior
region to contact the substratum. Upon settlement, the
larva placed the flattened, anterior region against the sub-
stratum, thereby assuming the definitive orientation of
the adult. Settled juveniles were not fixed to the substra-
tum and were capable of moving freely about using the
podia. During the first 10 days after settlement, juveniles
retained the ability to detach from the bottom and swim.
Well-developed juveniles have been obtained from the
plankton (F. S. Chia, pers. comm.).
Larvae that settled 2-3 days later than the majority
continued to develop at the same rate as the rest, even
though they remained planktonic. Likewise, juveniles that
resumed swimming after initial settlement continued to
develop at the same rate as those remaining on the bottom.
Therefore, settlement was not coupled to a rapid, drastic
metamorphosis into the juvenile form. In this respect,
Pteraster tesselatus is similar to other asteroids with pelagic
lecithotrophic larvae (e.g., Solaster endeca, Gemmill,
1912; Crossaster papposus, Gemmill. 1920) that undergo
an extensive but prolonged and gradual transformation
from larva to juvenile. But in contrast to most asteroid
larvae, settlement of Pteraster did not involve fixation to
the substratum nor a 90° bending (i.e.. flexion, sensu
Gemmill, 1912, p. 19) of the rudiment relative to the
larval body. In all other asteroid larvae, the disk of the
juvenile starfish lies in a sagittal plane in the posterior
part of the larval body, with the oral surface of the juvenile
on the left side of the larva (Fig. 4). Settlement involves
a bending of the larval body to bring the oral surface of
the juvenile disk against the substratum. In Pteraster, the
presumptive oral surface of the juvenile corresponds to
the yolky anterior region of the larva, which corresponds
to the animal pole of the embryo (site of polar body for-
mation). The aboral surface of the juvenile disk corre-
sponds to the posterior end of the larva and the vegetal
pole (blastopore) of the embryo (Fig. 4). Since the juvenile
disk lies in a transverse plane in the larval body, it does
not require flexion to attain the definitive orientation with
respect to the substratum following settlement.
Morphogenesis of the supradorsal membrane
Near the time of settlement, each of the five marginal
bulges developed a central indentation and became
strongly bilobed (Fig. 3B). Eventually they divided to yield
ten distinct marginal lobes around the juvenile (Fig. 5A,
B). Initially, these lobes were simple projections from the
larval surface. Later, they assumed a convoluted shape
and increased in size, nearly covering the lateral surface
of the aboral region of the juvenile disk (Fig. 5B). Simul-
LAHVAL ANTERIOR END
-LARVAL AXIS OF SYMMETRY
ADULT AXIS OF SYMMETRY
JUVENILE RUDIMENT
ADULT ORAL SURFACE
•ADULT ABORAL SURFACE
LARVAL RIGHT SIDE
ADULT AXIS OF SYMMETRY W
ADULT ABORAL SURFACE
LARVAL POSTERIOR END
LARVAL ANTERIOR END
LARVAL PLANE OF SYMMETRY
LARVAL LEFT SIDE
ADULT ORAL SURFACE
JUVENILE RUDIMENT
LARVAL POSTERIOR END
Figure 4. Diagrammatic representation of the location and orien-
tation of surfaces, planes of section, and axes of symmetry in the larval
body and rudiment of the juvenile disk of asteroids. Solid line represents
the anterior-posterior axis of the larva and the dashed line represents the
oral-aboral axis of the juvenile and adult. (A) Longitudinal view of the
larva of Pteraster tesselatus. (B) Ventral view of a generalized larva rep-
resentative of all other asteroids.
182
L. R. MCEDWARD
taneously with the splitting of the five marginal bulges to
yield ten marginal lobes, an additional five lobes formed
at the aboral pole of the juvenile (Fig. 5B). At this stage,
the juvenile consisted of an oral yolk plate directed to-
wards the substratum, a ring of podia located on the oral
surface of the disk (Fig. 5C), and a developing disk with
15 convoluted lobes on the aboral surface (10 marginal
and 5 aboral) (Fig. 5 A).
The aboral lobes developed differently in animals from
Vancouver Island compared to those from the San Juan
archipelago. In animals from the San Juan Islands, the
aboral lobes arose from the aboral regions of elongate
marginal bulges. The marginal bulges had a triangular
shape and gave rise to the three lobes (2 marginal + 1
aboral) from the vertices of the triangle. In contrast, an-
imals from Vancouver Island produced aboral lobes in-
dependently of the marginal lobes. A central stalk devel-
oped in the center of the aboral pole of the larva and
produced five rays. The tips of the rays became bulbous
and developed into aboral lobes.
Subsequent development of the lobes led to the for-
mation of the supradorsal membrane. The ten marginal
lobes became organized into five sets. The lobes in a set
were not derived from the same marginal bulge but instead
developed from neighboring bulges. Over a period of two
days, the lobes in each set enlarged and then fused along
their lateral edges. The aboral lobes fused along their lat-
eral edges and with the top (aboral) edge of the marginal
lobes. Later, fusion occurred among all of the lobes to
produce a complete secondary covering over the aboral
surface of the juvenile starfish (Fig. 6A). At the margin
of the disk, the oral edges of the marginal lobes extended
beyond the disk to form a skirt around the body. Fusion
along the lateral edges of the marginal lobes from adjacent
sets (i.e., between lobes derived from the same original
marginal bulge) was restricted to the aboral regions, leav-
ing a cleft extending between them from the oral edge of
the marginal skirt (Fig. 6A). This cleft was located above
the ambulacra! cluster of podia and marked the site where
the juvenile arms would form later in development. The
osculum was formed by the lack effusion along the central
edges of the five aboral lobes, which left a central opening
in the supradorsal membrane (Fig. 6B). Within ten days
of the start of lobe fusion, the supradorsal membrane was
completely smooth, without any visible indication that it
had formed from 1 5 separate elements. Ventilation of the
nidamental chamber by muscular pumping of the supra-
Figure 5. Scanning electron micrographs of newly settled juveniles
of Pteraster tesselalus. Scale bars = 0.2 mm. (Top) Lateral view. Age
= 16 days. Convoluted marginal lobes (ML) and aboral lobes (AL) cover
the aboral surface of the juvenile. Flattened yolky plate (Y) is located
below the circumferential groove (CG) and the podia (P). (Middle) Aboral
view. Age = 16 days. Ten marginal (ML) and five aboral lobes (AL)
cover the aboral surface. Podia (P) can be seen between some of the
marginal lobes and extending beyond the edge of the disk. (Bottom) Oral
view. Age = 12 days. The ten marginal lobes (ML) can be seen around
the edge of the disk of the juvenile. Podia (P) are arranged in a ring
around the yolk plate (Y).
A UNIQUE TYPE OF STARFISH LARVA
183
Figure 6. Scanning electron micrographs of juveniles of Pterasler
tessi'latus. Scale bars = 0.2 mm. (Top) Lateral view. Age = 27 days.
dorsal membrane (as occurs in adults) was not observed
in the juveniles, even six months after settlement (di-
ameter a; 2.0-2.1 mm). Ciliary activity was detected on
the external surface of the juvenile, including the supra-
dorsal membrane, but internal currents that might ven-
tilate the nidamental chamber could not be demonstrated
with dye streams. Mucus production was not obvious in
juveniles up to = 9 months after settlement (diameter
«= 2.1-2.3 mm).
Chronology of larval development
Larvae and juveniles were raised at ambient seawater
temperatures that ranged between 1 1-1 3°C. The schedule
of events during development of the larval and early ju-
venile stages is listed in Table 1. Development was re-
markably synchronous throughout all of the cultures and
among the larvae from different parents. However, the
age at settlement varied greatly. The majority (>80-95%)
of the larvae settled initially between days 10-12, but some
continued to swim (or resumed swimming) until the third
week. As indicated above, the schedule of development
was not influenced by the age at which settlement occurred
because the development of juvenile structures proceeded
normally in swimming larvae. Morphogenesis of the su-
pradorsal membrane was largely completed within 7-10
days following settlement. At the age of 1 month, the
juveniles measured =1.5-1.8 mm in diameter, and the
oral surface was still covered by a remnant of the yolk
from the anterior region of the larva (Fig. 6C). The juvenile
mouth had not yet formed and would not appear until
the end of the second month. There were no indications
of the juvenile arms. This resulted in a circular arrange-
ment of the podia (3-4 pair per ambulacral cluster) around
the oral surface of the disk. A radial arrangement of podia
is evident in most starfish larvae at the time of settlement
when the first two pair of podia are developing on the
juvenile arms (Fig. 7). Distinct arms were not present in
most Pteraster juveniles until the third month.
Discussion
Pattern of development in Pteraster tesselatus
The present study confirms the observation by Chia
(1966) that Pteraster tesselatus has free-swimming larvae
Aboral surface covered by the supradorsal membrane. Podia (P) extend
from under the oral side of the margin of the supradorsal membrane.
(Middle) Aboral view. Age = 27 days. Aboral surface covered by a com-
plete supradorsal membrane (SDM) formed from the fusion of the 15
lobes. The aboral body wall (BW) can be seen as the floor of the nida-
mental chamber through the large central osculum (O). (Bottom) Oral
view. Age = 28 days. Remnant of the yolk plate ( Y) lies in the center of
the disk. The mouth has not formed yet. The free edges of the marginal
lobes (ML) define the outer limit of the disk. Numerous podia (P) lie in
a circle around the yolk plate. Note that the juvenile arms have not
formed vet.
184
L. R. McEDWARD
and presumably pelagic development. Eggs were forcefully
expelled from the nidamental chamber through the os-
culum by the ventilatory flow. Eggs and embryos were
positively buoyant, as were larvae until close to the time
of settlement (8-10 days). Larvae were uniformly ciliated
and swam actively near the surface of the water, in the
laboratory. Later, larvae swam near the bottom and at-
tached to solid substrata, resulting in settlement from
plankton to benthos. Pteraster larvae and juveniles have
been obtained from the plankton in the San Juan archi-
pelago (F. S. Chia, pers. comm.; R. Emlet, pers. comnr.
R. Strathmann, pers. comm.).
My observations also confirm the lack of brooding in
this species. Chia (1966) did not find any brooded young
in the 12 animals that he dissected. I have dissected (i.e.,
opened or removed the supradorsal membrane) of >50
starfish without finding any evidence of brooding. Like-
wise, brooding has not been reported for this species by
any of the authors investigating other (nonreproductive)
aspects of Pteraster biology (Fisher, 1911, pp. 355-363;
Verrill, 1914, pp. 268-269; Rodenhouse and Guberlet,
1946; Mauzey et al, 1968; Johansen and Petersen, 1971;
Nance and Braithwaite, 1979, 1981).
Comparisons with previous descriptions of pterasterid
development
The only previous descriptions of morphological de-
velopment in pterasterids are a preliminary study of de-
velopment in Pteraster tesselatus by Chia (1966) and a
brief report on brooding in P. militaris (Kaufman, 1968).
My observations on P. tesselatus confirm many of Chia's
descriptions: egg size, egg color, jelly coat, wrinkled bias-
tula, large blastopore that later closes, ovoid early larva,
circumferential groove, anterior (= animal) and posterior
(= vegetal) body regions, resorption of the anterior region,
rudiment development in the posterior region, and the
chronology of early development (days 1-6).
However, there are also a number of substantial dif-
ferences. Chia reported that the arms of the juvenile star-
fish formed early in development (p. 508): "As soon as
the two parts [of the larval body] were clearly distinguish-
able, five primordial arms of the young seastar appeared
simultaneously in the vegetal part of the larval body and
the first pair of tube feet appeared on each arm (fig. 5)".
According to his chronological description, the arms were
formed by day 10 and the podia by day 13. I did not
observe arms even after one month (Fig. 6C). Probably
the initial five marginal bulges were misinterpreted as the
primordial juvenile arms. My observations show that the
marginal bulges are well developed by day 9, they are
arranged radially around the rudiment of the juvenile
disk, and they have clusters of podia associated with them
(Fig. 3B).
Figure 7. Light micrograph of a late-stage brachiolaria larva ofHen-
ricia sp. (Icviuxcula'.'). Scale bar = 0.2 mm. Lateral view from the left
side of the larva (= oral side of juvenile). Age = 28 days. Brachiolar arms
(BA) and attachment disk (AD) located at anterior end on the preoral
lobe (PL). Each juvenile arm (JA) has two pairs of developing podia (P)
and a terminal unpaired podium on the oral surface. The plane of larval
bilateral symmetry is parallel to the plane of the photograph.
I observed that most of the larvae settled between days
10-12. Chia reported that settlement occurred on the 25th
day. I attribute this difference to the conditions under
which the larvae were raised. A microbial and diatoma-
ceous film developed on the mesh bottoms of my culture
containers because they were suspended in unfiltered sea-
water and were not cleaned except when the larvae
hatched (day 3). Larvae settled readily on the filmed mesh,
starting on day 10. Both texture and film seem important
components of attractive surfaces for larval settlement in
Pteraster. Larvae never settled on the smooth, but filmed,
sides of the culture containers. Furthermore, larvae from
the same spawn that were held in closed culture contain-
ers, which were cleaned periodically, did not settle until
at least one week later than larvae in mesh-bottom con-
tainers. I have observed this difference in settlement with
the lecithotrophic larvae of other asteroid species that I
have cultured in the laboratory: Solaster stimpsoni, S.
dawsoni. S. endeca. Crossaster papposus, Henricia sp.
(leviusculaT). Chia cultured larvae of Pteraster tesselatus
in a small glass aquarium (pers. comm.) that probably
lacked an attractive surface (film or texture) for settlement.
Differences in the reported time of appearance of the
podia (2nd pair: day 25, Chia; day 9 present study) or the
number of podia present at a given stage can be explained
A UNIQUE TYPE OF STARFISH LARVA
185
by the difficulty of observing the podia; they develop deep
within the circumferential groove and are nearly hidden,
unless they are extended for exploration of the substratum
(Figs. 3B, 5A).
Other differences are less readily explained by culture
conditions or interpretation of structures. Chia reported
that the yolk mass was absorbed by day 25 and that the
mouth of the juvenile was open on day 30. A substantial
portion of the yolk plate was present on larvae in my
cultures on day 28, and the mouth did not open until a
month later. More puzzling is that Chia's description did
not mention the marginal or aboral lobes, nor the for-
mation of the supradorsal membrane. These were among
the most outstanding features of the development ofPtcr-
aster larvae in my cultures (Fig. 5 A, B). He does describe
the aboral epidermis as wrinkled and lacking spines or
ossicles. This suggests that the aboral surface was covered
by the supradorsal membrane because, in Pteraster tes-
selatus, the supradorsal membrane lacks ossicles. I ob-
served well-developed paxillae, by day 15, in juveniles
that had been cleared to make the supradorsal membrane
transparent. They should have been visible externally if
the supradorsal membrane had not yet formed in Chia's
cultures. Further, no mention was made by Chia of the
lack of brachiolar structures, the unusual symmetry of the
larva, or the orientation of the juvenile rudiment within
the larval body.
The striking differences between these two studies raise
interesting questions about development within the family
Pterasteridae. Development in Pteraster tesselatus may
be extremely variable, because I observed marked differ-
ences in the mode of aboral lobe formation in larvae from
different geographic regions. Alternatively, could we have
examined two different species? The systematics of the
genus Pteraster in the northeastern Pacific Ocean has not
been examined since the classic works of Fisher (191 1)
and Verrill (1914), and the family Pterasteridae is not
well known (M. Downey, pers. comm.). Clearly, the ex-
planation of the reported differences in development of
P. tesse/atus must await further examination of geographic
variation in development and analysis of specific and
subspecific systematics. Finally, Chia may have described
larvae undergoing abnormal development. He reported
that only 2% of the eggs in his cultures developed (Chia,
1966, p. 507). and the larvae in his Figures 3 and 5 do
not resemble the larvae that I observed. However, some
of these larvae successfully completed development into
apparently healthy juvenile starfish.
The development of only one other pterasterid has been
described. Pteraster militaris broods its young within the
nidamental chamber (Koren and Danielssen, 1856). Re-
cently it was shown that only a fraction of the eggs are
retained and brooded, the rest are spawned through the
osculum into the water column (McClary and Mladenov,
1988). Presumably the spawned eggs develop as pelagic
larvae, but their development has not been described.
Development of brooded embryos of P. militaris was
briefly described by Kaufman (1968). The embryos were
divided by a constriction (= circumferential groove) into
two hemispherical regions: an oral region consisting of
yolk that was eventually resorbed, and an aboral region
consisting of the rudiment of the body of the starfish. Five
radial (= marginal) bulges developed around the aboral
region. Subsequently, ten round tubercles (= marginal
lobes) developed from the five radial bulges. Podia de-
veloped between the tubercles. Later, five "arms" devel-
oped at the aboral end. Kaufman misinterpreted these
aboral "arms" as the arms of the juvenile starfish (see his
Fig. 2). They are probably equivalent to the five aboral
lobes of Pteraster tesselatus. and are therefore destined to
be incorporated into the supradorsal membrane. Surpris-
ingly, Kaufman did not mention the formation of the
supradorsal membrane. He stated that, when the juvenile
starfish was 1 .5-2.0 mm in diameter (7-10 days after for-
mation of the constriction), the madreporite opened near
the center of the aboral surface. I interpret this to be the
osculum of the supradorsal membrane. The young
emerged from the nidamental chamber of the mother
through transient slits in the supradorsal membrane.
P. militaris clearly broods at least some of its young.
However, if the embryos were artificially removed from
the brood chamber, they could swim via the uniform cil-
iation and could develop normally (Kaufman, 1968). The
free-swimming embryos became benthic at the same stage
that larvae of P. tesselatus settled. The similarity between
these two species suggests that, at least in some cases, there
are not major morphological differences between benthic
brooding and pelagic larval development within the Pter-
asteridae.
Unusual features of development in Pteraster tesselatus
The larval development of Pteraster tesselatus is very
different from what has been observed in other asteroids,
including velatids [e.g.. So/aster endeca (Gemmill, 1912),
Crossaster papossus (Gemmill, 1920), Solaster stimpsoni
and S. dawsoni (McEdward, unpubl. obs.)], and other
orders (see reviews by Hyman, 1955; Fell, 1967; Oguro
etai. 1988).
Morphogenesis of the supradorsal membrane. The most
striking feature of the development of Pteraster tesselatus
is the formation of 1 5 elaborate lobes on the external sur-
face of the posterior region of the larva (Fig. 5 A, B). This
feature alone makes the pelagic larva of P. tesselatus
unique among asteroids. The lobes fuse during the trans-
formation of the larva into the juvenile to produce the
supradorsal membrane (Fig. 6 A, B). It is not surprising
that a pterasterid larva has an unusual morphology, given
186
L. R. McEDWARD
that metamorphosis involves the formation of a highly
specialized structure found only in that family. Why the
supradorsal membrane forms so early in development,
before other juvenile features such as the arms and the
mouth, is not known.
Lack ofbrachiolar structures. All asteroids, except pax-
illosids, develop specialized larval attachment structures
consisting of the brachiolar arms and disk (Fig. 7) (Oguro
et ai, 1988). The brachiolar arms are for temporary at-
tachment to the bottom during exploration of the sub-
stratum, and the attachment disk is generally used to ce-
ment the settled larva to the substratum during meta-
morphosis (Barker, 1978). P. tesselatus, which is a velatid,
not a paxillosid, has a larva that does not form brachiolar
structures during development (Fig. 3A, B). Instead, podia
are used for attachment during settlement.
Accelerated development of the water vascular system.
Functional podia developed very early in Pteraster (day
7) compared to other species from the same geographic
region with pelagic lecithotrophic larval development (e.g.,
Solaster stimpsoni 25-30 days, 5. dawsoni =30 days,
Crossaster papposi is =50 days, Henricia sp.: 25-35 days;
McEdward, unpubl. obs.). Because Pteraster reproduces
in the summer and all of these other species develop in
the early spring, the rapid development of Pteraster might
be partly the result of higher seawater temperature (11-
13°C in summer vs. 7-9°C in spring). However, the podia
in Pteraster not only develop at a younger age, they are
also accelerated relative to other juvenile structures such
as the arms. In other asteroids with pelagic lecithotrophic
larval development, there are generally two pair of podia
and an unpaired terminal podium developing in the am-
bulacrum of each juvenile arm at settlement ( Fig. 7). These
podia do not become functional until days or weeks later.
In contrast, there were several pair of functional podia in
Pteraster long before the formation of the juvenile arms
(Table I; Fig. 6C). Heterochronic acceleration of podial
development in P. tesselatus provides a means of attaching
to solid substrata at settlement in the absence of the typical
larval attachment structures, the brachiolar apparatus.
Symmetry of the larva. The pelagic larva of Pteraster
is radially symmetrical, based on external morphology
(Fig. 3A, B) and the arrangement of internal structures
(e.g., coelomic cavities and water vascular system, Janies
and McEdward, in prep.). All other asteroid larvae that
have been described are bilaterally symmetrical (Fig. 7).
P. tesselatus and P. militaris are the only asteroids known
to lack a bilateral stage in the life cycle.
Larval and juvenile axes. The disk of the juvenile star-
fish lies in a transverse plane within the larval body of
Pteraster (Figs. 3B, 5 A) so that the anterior-posterior axis
of the larva is parallel to the oral-aboral axis of the juvenile
(Fig. 4). The anterior end of the larva corresponds to the
oral surface of the juvenile and the posterior end of the
larva becomes the aboral surface. The orientation of the
juvenile disk within the larva is very different in all other
starfish. Typically, the disk lies in a sagittal plane such
that the larval and juvenile axes are perpendicular and
the oral surface of the juvenile develops on the left side
of the larva and the aboral surface develops on the right
(Figs. 4, 7) (see review by Hyman, 1955).
Collectively, these features of morphology and devel-
opment distinguish the pelagic larva of P. tesselatus from
all other asteroid larvae. The distinctiveness of the Pter-
aster larva raises the question of its evolutionary origin.
It has been considered to be a lecithotrophic, modified
brachiolaria larva (Chia, 1966; Fell, 1967; Oguro et ai,
1988), presumably derived from an ancestor with pelagic
development. This interpretation is unlikely because it
requires modification of fundamental and highly conser-
vative features of larval morphology and metamorphosis
while those features were functional in the life cycle. As
an alternative explanation, the unusual larva of P. tesse-
latus might have evolved from an ancestor that brooded
its young (Strathmann, 1974). I suggest that most of the
unusual features (namely 1. 2, 4, and 5 from the list above)
can be interpreted as evidence of a highly derived mode
of development associated with the evolution of the spe-
cialized form of brooding in the pterasterids. The unique
structural modifications of adult pterasterids, which pro-
vide a brood chamber for the young, attest to the extreme
specializations towards brooding that have evolved in this
group. The evolution of this type of brooding probably
occurred during the radiation of the pterasterids in the
deep sea. Brooding development throughout much of the
evolutionary history of the pterasterids could have led to
the reduction and eventual loss of larval characteristics
in the offspring. For instance, the brachiolar structures
were probably lost in association with an entirely benthic
life cycle, where settlement structures are not needed. I
postulate that the degree of reduction of larval features
has been more extensive in pterasterids than in other as-
teroids. This does not require different selective forces
acting on pterasterid development, but rather could simply
be a function of differences in the evolutionary duration
of the brooding pattern of development in different taxa.
The relative stability of environmental conditions in the
deep sea over geological time scales could have resulted
in greater species longevity for pterasterid brooders com-
pared to brooding asteroids in other taxa in shallow water.
The result of extreme reduction and simplification would
be highly direct development from the embryonic stage
to the juvenile stage (McEdward, 1 989, and in prep.; Janies
and McEdward, 1991). This would explain nearly all of
the unusual features in the development of Pteraster tes-
selatus. Subsequent re-evolution of pelagic development,
probably in a shallow-water ancestor of P. tesselatus. re-
sulted in a larval form that was distinctly different from
A UNIQUE TYPE OF STARFISH LARVA
187
that of all other asteroids. Accelerated podial development
was probably a key event in the re-evolution of pelagic
development; it resulted in functional attachment struc-
tures that replaced the lost brachiolar complex and facil-
itated settlement to the benthos (McEdward, in prep.).
Acknowledgments
A. O. D. Willows, Director, provided space and facilities
at the Friday Harbor Laboratories, and J. Mclhnerney,
Director, provided space and facilities at the Bamfield
Marine Station. S. Carson and G. Gibson helped collect
starfish. SEM facilities were provided by the University
of Florida. Interdisciplinary Center for Biotechnology
Research and the Department of Zoology. A. Griffin, D.
Janies, and P. Eliazar assisted with preparation of speci-
mens for SEM and photography. J. Herrera prepared
specimens for clearing. Photographic printing was done
by D. Harrison. F.-S. Chia, M. Downey, J. Herrera, D.
Janies, R. Strathmann, and M. Strathmann discussed
ideas or reviewed various drafts of the manuscript. I thank
all of these colleagues for their assistance. Funding was
provided by the University of Florida Division of Spon-
sored Research (#89100245, #90012443, and DSR-B), the
Department of Zoology, University of Florida, the Friday
Harbor Laboratories, University of Washington, and the
Bamfield Marine Station, Western Canadian Marine Bio-
logical Society.
Literature Cited
Barker, M. F. 1978. Structure of the organs of attachment of brachio-
laria larvae of Stichaste r austral is (Verrill) and Coscinaslerias cala-
maria (Gray) (Echinodermata: Asteroidea). J. Exp Mar. Biol. EcoL
33: 1-36.
Blake, D. B. 1987. A classification and phytogeny of post-Paleozoic
sea stars (Asteroidea: Echinodermata). / Nat. Hist. 21: 481-528.
Chia, F.-S. 1966. Development of a deep-sea cushion star. Plerasier
tesselatus. Proc. Cai Acad. Sci. 34: 505-510.
Fell, H. B. 1967. Echinoderm ontogeny. Pp. S60-S85 in Treatise on
Invertebrate Paleontology. Part S. Echinodermata. R. C. Moore, ed.
Geological Society of America and Kansas Univ. Press, Lawrence,
Kansas.
Fisher, \V. K. 191 1. Asteroidea of the North Pacific and adjacent waters.
Part 1. Phanerozonia and Spinulosa. Smithsonian lust Bull. U. S.
Nat. Mus. 76:419 pp.
Fisher, VV. K. 1940. Asteroidea. Discover)- Rep 20: 69-306.
Gemmill, J. F. 1912. The development of the starfish Solaster endeca
Forbes. Trans. Zool. Soc. Land. 20: 1-71.
Gemmill, J. F. 1920. The development of the starfish Crossaster pap-
posus. Mullerand Troschel. Q J Microsc. Sci. 64: 155-190.
Hoeg, J. T. 1984. A culture system for rearing marine invertebrate
larvae and its application to larvae of rhizocephalan barnacles. J.
Exp. Mar. Biol. Ecol. 84: 167-172.
Hyman, L. H. 1955. The Invertebrates, Vol. 4, Echinodermata,
McGraw-Hill. New York. 763 pp.
Ives, J. E. 1888. On two new species of starfishes. Proc. Acad. Nat.
Sci. Phil 40:421-424.
Janies, D. A., and L. R. McEdward. 1991. Evolutionary significance
of a derived mode of coelom formation in the larva of the starfish.
Pteraster tesselatus. Am. Zool. 31: 105A.
Johansen, K., and J. A. Petersen. 1971. Gas exchange and active ven-
tilation in a starfish, Pleraster tesselalus. Z. vergl. Physiol. 71: 365-
381.
Kaufman, Z. S. 1968. The postembryonic period of development of
some White Sea starfish. Sov. J. Mar. Biol. 18: 507-510.
Koren, J., and D. C. Danielssen. 1856. Observations sur le developpe-
ment des asteries. Fauna Littoralis Norvegiae 2: 55-59.
Lambert, P. 1981. The Sea Stars of British Columbia. British Columbia
Prov. Mus., Victoria. 153 pp.
Mauzey, K. P., C. Birkeland, and P. K. Dayton. 1968. Feeding behavior
of asteroids and escape responses of their prey in the Puget Sound
region. Ecology 49: 603-619.
McClary, D. J., and P. V. Mladenov. 1988. Brood and broadcast: a
novel mode of reproduction in the sea star Pleraster militaris. Pp.
163-168 in Echinoderm Biology. R. D. Burke, P. V. Mladenov, P.
Lambert, and R. L. Parsley, eds. Balkema. Rotterdam.
McClary, D. J., and P. V. Mladenov. 1989. Reproductive pattern in
the brooding and broadcasting sea star Plerasier militaris. Mar. Biol.
103: 531-540.
McClary, D. J., and P. V. Mladenov. 1990. Brooding biology of the
sea star Pterasler militaris (O. F. Muller): energetic and histological
evidence for nutrient translocation to brooded juveniles. J. Exp. Mar.
Biol. Ecol. 142: 183-199.
McEdward, L. R. 1989. Development and evolution of a novel type
of starfish larva. Am. Zoo/. 29: 1 14A.
Nance, J. M., and L. F. Braithwaite. 1979. The function of mucous
secretions in the cushion star Pteraster tesselatus Ives. J. Exp. Afar.
Biol. Ecol. 40: 259-266.
Nance, J. M., and L. F. Braithwaite. 1981 . Respiratory water flow and
production of mucus in the cushion star, Pteraster tesselatus Ives
(Echinodermata: Asteroidea). J. Exp. Mar. Biol Ecol. 50: 21-31.
Nation, J. L. 1983. A new method using hexamethyldisilazane for
preparation of soft insect tissue for scanning electron microscopy.
Slain Techno/. 58: 347-351.
Oguro, C., M. Komatsu, and Y. T. Kano. 1988. Significance of the
nonbrachiolarian type of development in sea-stars. Pp. 241-246 in
Echinoderm Biology. R. D. Burke, P. V. Mladenov, P. Lambert, and
R. L. Parsley, eds. Balkema, Rotterdam.
Rodenhouse, I. Z., and J. E. Guberlet. 1946. The morphology and
behavior of the cushion star Pteraster tesselatus Ives. Univ. of Wash-
ington. Publ. Biol. 12: 21-48.
Sladen, W. P. 1889. Report on the Asteroidea collected by the H.M.S.
Challenger. Challenger Rep.. Zool. 30: 888 pp.
Strathmann, M. F. 1987. Reproduction and Development of Marine
Invertebrates of the Northern Pacific Coast. Univ. Washington Press.
Seattle. Washington. 670 pp.
Strathmann, R. R. 1974. Introduction to function and adaptation in
echinoderm larvae. Thai. Jugoslav. 10: 321-339.
Strathmann, R. R. 1978. The evolution and loss of feeding larval stages
of marine invertebrates. Evolution 32: 894-906.
Verrill, A. E. 1914. Monograph of the shallow-water starfishes of the
North Pacific coast from the Arctic Ocean to California. Smithsonian
Inst.. Harriman Alaska Series 14: 408 pp.
Reference: Biol. Bull- 182: 188-194. (April, 1992)
Age of the Mangrove Crab Scylla sermta
at Colonization by Stalked Barnacles
of the Genus Octolasmis
WILLIAM B. JEFFRIES', HAROLD K. VORIS2, AND SOMBAT POOVACHIRANON3
^Department of Biology, Dickinson College, Carlisle, Pennsylvania 17013; 2 Department of Zoology,
Field Museum of Natural History, Chicago, Illinois 60605; and ^Phuket Marine Biological Center,
Phuket 83000, Thailand
Abstract. Cyprid larvae of the lepadomorph Octolasmis
colonize the gill chambers of the edible mangrove crab,
Scylla serrata (Forskal, 1755), sometimes in debilitating
numbers. We set out to determine when, in the life cycle
of the host, barnacle infestation begins. A total of 856
mangrove crabs, ranging in size from 10.9 to 132.3 mm
carapace width (instars 5 to 18), were collected from nat-
ural populations in Phuket, Thailand, and examined for
these barnacles. Almost a third harbored one or more
barnacles. The smallest crab to host a barnacle was 34.3
mm (instar 10); 233 smaller crabs, representing instars 5-
9, had none. Infestations by more than one barnacle were
uncommon among crabs of less than 70 mm carapace
width (instar 13). The percentage of crabs hosting bar-
nacles increased as the crabs approached sexual maturity,
and the magnitude of infestation on individual crabs in-
creased with their size. The distribution of octolasmids
on the gills of immature crabs differed from that on mature
crabs. In the former, all barnacles were on the inside of
the gill surfaces and none were on the outside, whereas
in the latter, 1 1% were on the outside of the gills. The
numbers of barnacles on the inside and the outside of the
gills is a function of the number of barnacles in the gill
chamber. The major inhalant aperture size, and gill
chamber size were eliminated as possible factors limiting
infestation. Instars 10 and 1 1 may be suboptimal for in-
festation by octolasmids because the intermolt time be-
tween instars does not allow sufficient time for production
of barnacle nauplii. Current data do not permit us to dis-
tinguish the relative influences of microhabitat use, host
hormonal changes, and behavioral changes on infestation.
Received 16 September 1 99 1 ; accepted 12 December 1991.
Introduction
Many interesting and diverse symbiotic relationships
(sensu lato, de Bary, 1879) exist between decapods and
metazoans of different phyla. The growth of some rhi-
zocephalans, e.g., Sacculina spp., results in parasitic cas-
tration of hosts so that male crabs develop some of the
secondary sexual and behavioral characteristics of female
crabs (Bang, 1983; Cressey, 1983; Overstreet, 1983). In-
festation of crab branchial lamellae and subsequently crab
egg masses by nemertean worms, e.g., Carcinomemertes
spp., likely impedes the free flow of water over the gills
and results in the predation of host eggs (Humes, 1942).
Colonization of crab respiratory surfaces by Octolasmis
spp. may result in heavy infestations that overwhelm the
cleaning capacity of grooming appendages and make res-
piration difficult (Overstreet, 1983).
As is the case with many edible crab species, the respi-
ratory chambers of the mangrove crab, Scylla serrata
(Forskal, 1755), are inhabited by stalked barnacles, which
occupy space on the gills normally available for respiratory
exchange of oxygen and carbon dioxide. To colonize the
crabs, octolasmid cyprid larvae collect on the host just
prior to molt and transfer from the old exoskeleton to the
newly molted crab at the time of ecdysis (Jeffries et al.,
1989). Following attachment, the cyprid metamorphoses
to the adult barnacle body form and is sessile thereafter.
Because they are sessile, these barnacles cannot recolonize
crabs that have molted. Thus for a barnacle to achieve
reproductive success, there must be a sufficient interval
between crab molts for the cyprid to attach, metamor-
phose to the adult form, reach sexual maturity, copulate,
oviposit, and release nauplii.
188
COLONIZATION OF MANGROVE CRABS
189
CRAB SIZE (carapace width, mm)
120 100 80 60 40 20
MALE CRABS
(n = 403)
40
60 80 100 120
FEMALE CRABS
(n = 453)
Figure 1. The size distributions for 403 male and 453 female man-
grove crabs (Scylla serrata) collected from mangrove areas in the im-
mediate vicinity of Phuket. Thailand.
Some small crabs also harbor octolasmids, e.g., Uca
minax, with a carapace width of 38 mm (Williams, 1965)
was reported to bear Octolasmis miilleri (Coker, 1902),
(Pearse, 1936). In a survey of the decapods of Sabah, Ma-
laysia (unpublished), we noted that three different crab
species, all small, harbored Octolasmis angulata in their
branchial chambers: 2 male Pilumnits scabriusculus
(Adams & White), 32.1 and 36.3 mm, respectively; 1 male
Pilumnits vespertilio (Fabricius, 1793), 23.2 mm; and 1
male Actaeodes sp., 26.9 mm. These observations, to-
gether with the observation that smaller S. serrata had
few or no octolasmids. prompted us to ask whether host
size might be a consideration in the mangrove crab-oc-
tolasmid symbiosis.
The purpose of this research was to determine the stage
and time in the life cycle of the mangrove crab when
Octolasmis spp. colonize its gill chambers. Specifically,
we sought to identify the youngest crab instars that harbor
Octolasmis. and to compare the number and distribution
of barnacles on their gills with those on mature crabs.
Materials and Methods
During 1990 and 1991, S. serrata were collected from
shallow seas adjacent to mangroves, mostly within 2 km
of the town of Phuket, Thailand, for study at the Phuket
Marine Biological Center. The very small instars were
caught by hand, whereas larger crabs were caught in baited
traps. The crabs were sexed, weighed, and their carapace
lengths and widths measured (Heasman, 1980). They were
preserved in formalin and stored in 70% ethanol for later
examination. The crabs were examined for Octolasmis
cyprids, juveniles, adults with distinct ovaries, and gravid
adults. The exact location (left or right gill chamber, gill
number, inside or outside gill surface, proximal, medial,
or distal region of gill), and the length of the capitulum
of each barnacle were recorded using the methods pre-
viously employed (Jeffries el a!., 1982). A dissecting mi-
croscope was used to determine the reproductive status
of the barnacles.
Results
A total of 856 S. serrata were examined; 403 were males
and 453 females. The carapace width of the male and
female crabs ranged from 10.9 to 125.5 mm and from
1 1.2 to 132.3 mm, respectively. The size distributions of
the male and female crabs examined were very similar
(Fig. 1).
Of the 6648 barnacles observed, 168 were cyprids, 3670
were Octolasmis cor, 1758 were Octolasmis angidata, and
the remainder were too small to identify. Except where
noted, all stages and both species were pooled in the anal-
ysis because the focus of this paper is on barnacle colo-
nization relative to crab age. In a subsequent paper about
the ecology of barnacles resident on mangrove crabs, we
will address the differences between the two species. That
subject deserves individual treatment because opinion is
divided on whether S. serrata branchial chambers bear
several varieties of Octolasmis cor (Monod, 1922; New-
man, 1 960) or two species, O. cor and O. angulata, as we
assert.
The percentage of crabs that harbored barnacles was
very low (<5%) for crabs of less than 50 mm carapace
width, whereas the incidence of infestation rose sharply
as crab size increased above 50 mm (Fig. 2).
CRAB SIZE (carapace width, mm)
1 - 10
11-20
21 - 30
31 - 40
41 - 50
51 -60
61 - 70
71 - 80
81 - 90
91 - 100
101 - 110
111 - 120
121 - 130
10 20 30 40 50 60 70 80
PERCENT CRABS COLONIZED
90 100
Figure 2. The percentage of mangrove crabs (Scylla serrata) of dif-
ferent sizes with one or more octolasmid barnacles present in a gill
chamber. The 856 crabs range in size from I0.9to 1 32. 3 mm in carapace
width.
190
W. B. JEFFRIES ET AL.
Of the 856 crabs examined for barnacles, 260 (30.4%)
had one or more barnacles in their gill chambers. Of these,
134 crabs were males and 126 were females. The per-
centage of male crabs harboring barnacles (33.3%) was
only slightly greater than for female crabs (27.8%), and
both males and females showed increased numbers of
barnacles with increased carapace width (Fig. 3).
The smallest crab in the sample bearing a barnacle in
its gill chamber was a male with a carapace width of only
34.3 mm (instar number 10; Ong, 1966). The solitary
octolasmid (O. sp.) was found on the inside of the seventh
gill; it was reproductively immature and had a capitular
length of 0.86 mm. The next smallest crab to have a bar-
nacle in its gill chamber was a female with a carapace
width of 43.1 mm (instar 1 1; Ong, 1966). This crab had
a single cyprid on the inside of the sixth gill. The carapace
widths of all other barnacle-bearing crabs were more than
50.0 mm, corresponding to instar 1 2 or greater.
Two male crabs with carapace widths of 5 1 .3 mm were
the next smallest crabs to harbor single barnacles. One
crab had a cyprid, whereas the other was the smallest crab
to possess a sexually mature Octolasmis cor, measuring
1.72 mm in capitular length, with a distinct ovary indi-
cating that it was potentially reproductively active.
The smallest crab with multiple barnacles, a female,
had a carapace width of 54.5 mm; it harbored two O. cor
with distinct ovaries and capitular lengths of 1.86 and
2.43 mm. Figure 2 shows that the incidence of barnacles
increased on crabs between 60 and 70 mm, but most crabs
below 70 mm had no more than 5 barnacles. One notable
exception was a female crab with a carapace width of 60.4
mm and 95 barnacles. This crab was also the smallest
crab to harbor ten or more barnacles (Fig. 3). The 95
barnacles ranged in size from 0.57 to 2.57 mm in capitular
length (X = 1 .66 mm). Of the 95 barnacles, 38 had distinct
ovaries, and the smallest of these had a capitular length
of 1.43 mm. Eleven were gravid with ovigerous lamellae,
and the smallest of these was 1 .72 mm in capitular length.
Of five bearing nauplius I larvae about to hatch (stage N;
Lewis, 1975), the smallest was also 1.72 mm in capitular
length. This crab was exceptionally small for the level of
infestation, and the next smallest crab with 10 or more
barnacles had only 1 1; its carapace width was 67.3 mm.
The distribution of barnacles (n = 6648) between the
left and right chambers did not differ significantly from
50:50 (Binomial test, P < .01), and thus the left and right
chambers were pooled for the following distribution com-
parisons.
To explore possible differences in patterns of coloni-
zation among crabs of different sizes, the 260 infected
crabs were divided into two groups: those with carapace
widths below 70 mm (n = 87) and those above 70 mm
(n = 173). This division point was selected on the basis
of observed infestation levels (Fig. 3) and the knowledge
that 70 mm (instar 14) likely corresponds to the beginning
of crab sexual maturity. Among female mangrove crabs
of instar 14, the ovaries are considerably developed and
easy to see grossly as was reported for female Ca/linectes
sapidus (Rathbun, 1896) following the penultimate molt
(Johnson, 1980).
All 303 barnacles found on the 87 crabs with carapace
widths of less than 70 mm were located on the inside
surface of the gills (Table I). This distribution is signifi-
cantly different (df = 1. X2 = 38.8, P < .001) from that
350
300
(3 300
<H
O
u 150
E
3
Z 100
so
30
50
90
Crab Carapace Width (mm)
110
130
Figure 3. The relationship between the crab carapace width and the number of octolasmid barnacles
present in the gill chambers for 134 male and 126 female mangrove crabs (Scylla serraia).
COLONIZATION OF MANGROVE CRABS
Table I
The distribution of barnacles over the inside and outside surfaces of the gills of the 87 crabs with carapace widths less than 70 mm
191
Gill
number
Inside of gills
Outside of gills
In + Out
totals
Proximal
Medial
Distal
Totals
Proximal
Medial
Distal
Totals
1
1
4
0
5
0
0
0
0
5
2
0
0
0
0
0
0
0
0
0
3
9
22
0
31
0
0
0
0
31
4
33
32
1
66
0
0
0
0
66
5
27
38
1
66
0
0
(1
0
66
6
26
42
5
73
0
0
0
0
73
7
16
31
2
49
0
0
0
0
49
8
5
8
0
13
0
0
0
0
13
Totals =
117
177
9
303
0
0
0
0
303
On Rakers
1
Total =
304
observed for the 173 crabs with carapace widths of greater
than 70 mm, where 5609 barnacles were found on the
inside surface of the gills and 723 (11%) were found on
the outside (Table II). The smallest crab having a barnacle
attached to the outside surface of its gills had a carapace
width of 76.8 mm, and the inside surface of the gills in
that chamber had a total of 20 barnacles. The smallest
crab having multiple barnacles attached to the outside
surface of its gills had a carapace width of 78.2 mm, and
the numbers of barnacles on the inside and outside sur-
faces of the gills were 64 and 2 in the left chamber, and
6 1 and 3 in the right chamber.
For the 87 crabs less than 70 mm, and the 173 above
70 mm in carapace width, the distribution of barnacles
on the inside of gills 1 through 8 (Tables I, II) did not
differ significantly (df = 7, X2 = 6.1, P > .05). Nor did
the distribution of barnacles along the length of the gills
(proximal, medial, and distal) on the inside surface differ
significantly (df = 2, X2 = 2.4, P > .05) between the crabs
above and below 70 mm.
The distribution of barnacles on the inside versus the
outside of the gills among all 260 crabs with barnacles
(data from Tables I and II combined) showed significant
differences. The distribution of barnacles on the inside
versus the outside, and over gills 1 through 8, differed
significantly (df = 7, X2 = 525.9, P < .001). In addition,
the distribution of barnacles along the length of the gills
(proximal, medial, and distal) on the inside versus the
outside surface differed significantly (df = 2, X2 = 87.3,
P > .001). Whether on the inside or outside surface of
the gills, barnacles in this sample were least common on
the distal third of the gills. On the inside surface, the me-
dial section of the gill was the most densely populated,
whereas on the outside surface, the proximal portion of
the gills was the most populated.
Discussion
These data support four major conclusions about col-
onization of mangrove crabs by barnacles: (1) juvenile
crabs are virtually free of octolasmids; (2) as crabs ap-
proach sexual maturity, the percentage of crabs hosting
octolasmids increases; (3) the level of barnacle infestation
of individual crabs increases with crab size; and (4) the
distribution of octolasmids on the gills of immature crabs
is different from that on mature crabs. Our discussion will
focus on possible explanations for these findings.
In this study we found that 233 crabs, representing in-
stars 5 through 9 (Ong, 1966), did not host a single bar-
nacle. Several previous studies of decapods and their
symbionts also have suggested size thresholds for infes-
tation. Although the intertidal shore crab, Hemigrapsus
oregonensis (Dana, 1851), of either sex, mature or not,
may become infested with juvenile nemertean worms,
Carcinonemertes epialti Coe, 1902, a threshold carapace
width of 8 mm exists; above that threshold, both the in-
cidence of infestation and the average burden of worms
increase dramatically with increasing host size (Kuris,
1 978). Also, juveniles of the edible dungeness crab, Cancer
magister, below 20 mm carapace width are not infested
by Carcinonemertes errans. whereas worm burdens in-
crease with crabs above 20 mm (Wickham, 1980). Me-
tacercariae of the trematode, Spelotrema excellens Nicoll,
were found in 103 of 1 1 5 specimens of the portunid crab,
Carcinus maenas (L.). Twelve crabs with carapace widths
of less than 15 mm were not parasitized, whereas all above
1 5 mm were parasitized, and there was a significant cor-
relation between the crab carapace width and the intensity
of infection (Threlfall, 1968).
These studies suggest that size thresholds do exist among
the host decapods in such symbiotic relationships. How-
ever, there has been very little critical examination of the
192 W. B. JEFFRIES ET AL.
Table II
The distribution of barnacles over the inside and outside surfaces of the gills of the 1 73 crabs with carapace widths greater than 70 mm
Gill
number
Inside of gills
Outside of gills
In + Out
totals
Proximal
Medial
Distal
Totals
Proximal
Medial
Distal
Totals
1
37
36
4
77
1
1
23
25
102
2
5
4
6
15
14
13
9
36
51
3
162
398
45
605
16
18
2
36
641
4
607
691
33
1331
87
41
0
128
1459
5
431
564
39
1034
209
95
10
314
1348
6
579
577
43
1199
72
73
2
147
1346
7
322
633
66
1021
11
18
2
31
1052
8
144
171
12
327
0
6
0
6
333
Totals =
2287
3074
248
5609
410
265
48
723
6332
On Rakers
12
Total =
6344
possible effects of host ontogenetic changes on symbiosis
involving Crustacea. This is important because the way
that symbionts interact with different host ontogenetic
stages can contribute to our understanding of the mech-
anisms underlying the symbiotic relationships.
For the relationship between S. serrata and Octolasmis
spp., we have considered the following possible influences
on the observed non-random distributions: ontogenetic
differences in physical barriers such as the size of the major
incurrent respiratory apertures, and gill chamber size; du-
ration of intermolt period; crab macrohabitat distribution;
crab microhabitat distribution; and crab behavior.
In order to be a limiting factor, the incurrent respiratory
apertures of the crab would have to be small enough to
exclude cyprid larvae. The average length and diameter
of 10 preserved Octolasmis cyprids was 0.82 X 0.35 mm.
By comparison, the average width and height of the major
apertures of five of the smallest crabs (carapace width 10-
1 9 mm) examined in this study was 1 .9 X 1 .0 mm. Clearly,
crab intake aperture size is not a limiting factor to cyprid
larvae trying to enter the branchial chambers of S. serrata.
If space in the gill chamber were a major factor in lim-
iting the occurrence of barnacles on immature crabs, we
would expect the numbers of barnacles on crabs to in-
crease steadily with increased crab size. This does not seem
to be the case. For example, two crabs with carapace
widths of 60.4 and 61.4 mm had large populations of
barnacles, although their sizes are at the threshold where
colonization is just beginning (Fig. 3).
For the length of the crab intermolt period to strictly
limit the effective colonization by barnacles, the crab in-
termolt period must be shorter than the total time nec-
essary for the cyprid to attach, metamorphose, reach sex-
ual maturity, become gravid, and release nauplius I larvae.
This is the case because post-metamorphic barnacles are
sessile, and when a crab molts, individual barnacles re-
main attached to the crab exuviae and are unable to re-
colonize a host. Thus there is the potential for a fine-
tuned host/symbiont relationship predicated on the in-
termolt time period of the host and the time the symbiont
requires to complete its life cycle.
Earlier work provided an estimate of 24 h for Octolas-
mis cyprid attachment to metamorphosis, an estimated
daily growth rate of 0.336 mm in capitular length (Jeffries
el a/., 1985), and demonstrated that the major coloniza-
tion of crabs by cyprids occurs immediately after ecdysis
(Jeffries rt al., 1989). In this study, the capitular length
of the smallest Octolasmis cor (1.72 mm) observed with
ovigerous lamellae compares favorably with the 1 .6 mm
specimen reported with mature eggs by Matheswari and
Fernando ( 1989), and exceeds the 1.14 mm gravid spec-
imen of O. miilleri reported by Jeffries and Voris ( 1983).
On average, the capitular length of the newly metamor-
phosed barnacle is 0.57 mm, and thus it takes about 3.4
days to reach 1.72 mm at the daily growth rate of 0.336
mm. Hence, the time from cyprid attachment to a gravid
barnacle at oviposition with ovigerous lamellae is about
4.4 days.
The development time from oviposition to the release
of nauplius I larvae is unpublished for octolasmids, but
it has been reported for other lepadomorphs. For Ibla
quadrivalvis Cuv., it was 16 to 17 days at 23°C (Ander-
son, 1964), and for Pollicipes polymerus, it averaged 25.4
days at 12 to 16°C (Lewis, 1975). Because seawater tem-
peratures at the S. serrata collection sites near Phuket,
Thailand, range from 27 to 31°C, it is very likely that
for O. cor the required time from oviposition to release
of nauplius I larvae is no more than 14 days, and may
be as little as one week. Thus, the time required for O.
cor to attach and reproduce is likely no more than 18.4
days (4.4 plus 14) and may be as little as 1 1.4 days (4.4
plus 7).
COLONIZATION OF MANGROVE CRABS
193
200 r
J3 ISO
u
a
_
«100
a
m
50
30
50
70 90 110
Crab Carapace Width (mm)
130
Figure 4. Of the 260 crabs with barnacles, 435 of the 520 gill chambers contained one or more barnacles.
For each chamber, the size of the crab is plotted against the number of barnacles present in the chamber.
The symbol size indicates the percentage of the barnacles present on the outside surface of the gills in
increments of 10%.
S. serrata instar 9 crabs range in size from 26 to 33
mm in carapace width and have an intermolt period of
at least 1 5 days (Ong, 1966). The absence of any barnacles
on the 233 crabs of instars 5 to 9 in this study may be
attributable to barnacle avoidance because the intermolt
period is too short to allow for barnacle reproduction.
Crab instars 10 to 1 1 range from 33 to 48 mm in car-
apace width and have minimum intermolt periods of 16
and 22 days, respectively. According to our earlier cal-
culations, these are minimum time periods that octolas-
mids require to produce offspring and they would not
allow individuals to produce multiple broods, thus lim-
iting fecundity (Jeffries et a/., 1985). In this study, the fact
that only a small fraction of instars 10 and 1 1 harbored
barnacles (Fig. 2) and none had heavy infestations (Fig.
3) is consistent with the idea that these crab instars rep-
resent suboptimal substrata.
Instars 1 2 and above have carapace widths greater than
45 mm and intermolt periods of 30 days or more. Such
intermolt periods are sufficient to allow for the production
of multiple broods. This is consistent with our observa-
tions that the incidence of infestation increases among
crabs above 50 mm carapace width (Fig. 2) and is further
supported by the observation that one crab of 60.4 mm
carapace width (instar 13) had 95 barnacles (Fig. 3), many
of which harbored nauplius I larvae.
In this study, mangrove crabs of different sizes — ju-
veniles and adults — were collected at the same time in
the same shallow water mangrove macrohabitat, thus di-
minishing the possibility of a difference in availability of
cyprids to potential hosts at the macrohabitat level.
Differences in infestation rates among successive instars
of S. serrata might be the result of ontogenetic changes
in microhabitat use, shifts in host hormone levels (che-
motaxis), or behavioral changes. However, we cannot dis-
criminate among these factors on the basis of current in-
formation. For example, it was reported that juvenile
mangrove crabs use shelter under existing bottom debris
in open areas (Heasman, 1980), whereas older crabs use
burrows. But we do not know whether this difference has
an impact on mangrove crab colonization by Octolasmis
cyprid larvae.
The differences observed in the distribution of barnacles
on the inside and outside of the gills in crabs less than
and those greater than 70 mm carapace width (Tables I,
II) could be due to a density-dependent response to
crowding on the inside surface of the gills, or it could be
due to a lack of space in the chamber on the outside surface
of the gills. The former explanation is better supported
by our data. If space on the outside of the gills controlled
colonization, we would expect larger crabs with modest
numbers of barnacles (e.g., 5 to 20) to have barnacles on
both the inside and outside surface of their gills. This is
not the case. Barnacles are only found on the outside gill
surfaces in significant numbers when the inside surface
has 20 or more barnacles, regardless of crab size (Fig. 4).
194
W. B. JEFFRIES ET AL.
These non-random, species-specific distributions suggest
the next generation of research questions to be addressed.
Acknowledgments
We thank the staff of the Phuket Marine Biological
Center for their logistical support and the use of their
facilities. In particular, we are grateful to the Director,
Mr. Udom Bhatia, for his generous cooperation. Special
thanks are due to Mr. Boonchoy Kuoyratanakul, a fish-
erman who collected most of the crabs and Mr. Saengdee
Chailert who received crabs from fishermen for us. We
thank Helen Voris for numerous helpful editorial com-
ments. Support from the Dickinson College Board of Ad-
visors and the Faculty Research Fund, and the Field Mu-
seum of Natural History Research Fund made this in-
vestigation possible.
Literature Cited
Anderson, D. T. 1965. Embryonic and larval development and segment
formation in Ibla quadrivalvis Cuv. (Cirripedia). Aust. J ZooL 13:
1-15.
Bang, F. B. 1983. Crustacean disease responses. Pp. 1 13-153 in The
Biology of the Crustacea. Vol. 6, Anthony J. Provenzano, Jr., ed.
Academic Press. Inc. New York.
Cressey, R. F. 1983. Crustaceans as parasites of other organisms. Pp.
251-273 in The Biology of the Crustacea. Vol. 6, Anthony J. Prov-
enzano. Jr., ed. Academic Press, Inc. New York.
de Bary, A. 1879. Die Erscheinung Der Symbiose, Karl J. Trubner.
Strassburg.
Heasman, M. P. 1980. Aspects of the general biology and fishery of
the mud crab Scylla serrata (Forskal) in Moreton Bay, Queensland.
Ph.D. Thesis No. 2210, University of Queensland. 506 pp.
Humes, A. G. 1942. The morphology, taxonomy, and bionomics of
the nemertean genus Carcinonemertes. Illinois Biol. Monogr. 18: 1-
105.
Jeffries, \V. B., and H. K. Voris. 1983. The distribution, size and re-
production of the pedunculate barnacle, Octalasmis miilleri (Coker,
1902) on the blue crab, Callinectes sapidus (Rathbun, 1896). Fiel-
diana Zoo/. N. S 16: 1-10.
Jeffries, W. B., H. K. Voris, and C. M. Yang. 1982. Diversity and
distribution of the pedunculate barnacle Octolasmis in the seas ad-
jacent to Singapore. / Crustacean Biol 2(4): 562-569.
Jeffries, W. B., H. K. Voris, and C. M. Yang. 1985. Growth of Oc-
tolasmis cor (Aurivillius. 1 892) on the gills of Scylla serrata (Forskal,
1755). Biol. Bull. 169: 291-296.
Jeffries, W. B., H. K. Voris, and C. M. Yang. 1989. A new mechanism
of host colonization: pedunculate barnacles of the genus Octolasmis
on the mangrove crab Scylla serrata. Ophelia 31: 51-58.
Johnson, P. T. 1980. Histology of the blue crab, Callinectes sapidus.
Praeger Publishers, New York.
Kuris, A. M. 1978. Life cycle, distribution and abundance of Carci-
nonemertes epialti. a nemertean egg predator of the shore crab, Hem-
igrapsus oregonensis. in relation to host size, reproduction and molt
cycle. Biol. Bull 154: 121-137.
Lewis, C. A. 1975. Development of the gooseneck barnacle Pollicipes
polymerus (Cirripedia: Lepadomorpha): fertilization through settle-
ment. Mar. Biol. 32: 141-153.
Matheswari, R., and S. A. Fernando. 1989. Fecundity of epizoic bar-
nacles. Indian J Inv. Zoo/. Aqua. Biol 1(2): 31-35.
Monod, T. 1922. Sur un Dichelaspis de Madagascar, commensal de
Scylla serrata (Forskal). Extrait du Bulletin de la Societe Zoologique
de France XLVI: 264-269.
Newman, W. A. 1960. Five pedunculate cirripeds from the western
pacific, including two new forms. Crustaceana 1: 100- 1 16.
Ong, K. S. 1966. Observations on the post-larval life history of Scylla
serrata. Forskal, reared in the laboratory. Malays. Agric. J. 45(4):
429-443.
Overstreet, R. M. 1983. Metazoan Symbionts of Crustaceans. Pp. 155-
250 in The Biology of Crustacea. Vol. 6. Anthony J. Provenzano,
Jr.. ed. Academic Press, Inc. New York.
Pearse, A. S. 1936. Estuarine animals at Beaufort, North Carolina. J.
E/isha Mitchell Sci. Soc. 52(2): 174-222.
Threlfall, W. 1968. Note on metacercariae of Spelotrema excellens
Nicoll in Carcinus maenas (L.). / Exp Mar Biol. Ecol. 1: 1 54- 1 55.
Wickham, D. E. 1980. Aspects of the life history of Carcinonemertes
errans (Nemertea: Carcmonemertidae). an egg predator of the crab
Cancer magister. Biol Bull 159: 247-257.
Williams, A. B. 1965. Marine decapod crustaceans of the Carolinas.
Fish. Bull. U. S. Fish Wildlife Serv. 1: 1-298.
Reference: Biol. Bull 182: 195-209. (April, 1992)
Architectural and Mechanical Properties of the Black
Coral Skeleton (Coelenterata: Antipatharia):
A Comparison of Two Species
KIHO KIM*. WALTER M. GOLDBERG**, AND GEORGE T. TAYLOR
Florida International University, Department of Biological Sciences,
University Park, Miami, Florida 33199
Abstract. Black coral skeletons are laminated compos-
ites, composed primarily of chitin fibrils and non-fibrillar
protein. This paper examines mechanical properties of
the composite and the architecture of the chitin compo-
nent. Two species are shown to differ significantly in their
tensile strength and fibril structure. The skeleton of An-
tipathes salix, a Caribbean species of commercial value,
is stiffer, harder, darker, more dense, and more hydro-
phobic than Antipathesfiordensis from New Zealand. The
chitin fibrils constitute a greater proportion of the skeleton
in A. salix. where they are helically wound in an anti-
clockwise pattern within layer. Adjacent layers of skeleton
are arranged with relatively small layer-to-layer fibril
biases. There is no evidence of "helicoidal" structure in
this skeleton. The fibrils in A. ftordensis are also wound
anticlockwise within layer, but with rather large fibril
biases between layers, giving the appearance of a mesh-
work. Large-scale helicoidal patterns with apparent ro-
tations of 180° characterize this material. Skeletal archi-
tecture is compared with the cuticle of insects and other
arthropods. The skeletons of both species exhibit spines
characteristic of the Antipatharia. We suggest that these
have a significant reinforcing effect on the strength of the
skeleton, contributing to an overdesign for the habitat in
which these organisms presently occur.
Introduction
Black corals are well known as articles of commerce in
the jewelry and curio trade dating from at least the time
Received 21 October 1991; accepted 24 January 1992.
* Current address: State University of New York at Buffalo,
Department of Biological Sciences. Buffalo, New York, 14260.
** To whom correspondence should be addressed.
of the ancient Greeks (Hickson, 1924). The skeleton can
be polished to an onyx-like luster, but because it is organic,
it can also be bent and molded while being heated. Con-
sequently, this material is highly prized in the jewelry
trade, and collection pressure has resulted in the listing
of black corals in the Convention for the International
Trade in Endangered Species (CITES) (Wood and Wells,
1988). Collection pressure notwithstanding, certain species
of black coral may constitute a substantial part of the
fauna on deep reefs and in other more geographically re-
stricted habitats. In the Fiordland district on the southwest
coast of New Zealand's south island, for example, the
endemic black coral, Antipathesfiordensis Grange is the
dominant macrobenthic organism in depths of 10-20 m
(Grange. 1985, 1990). Similarly, Antipathes salix Pour-
tales, is one of the more commonly occurring black coral
species in the Caribbean, although it is most often re-
stricted to coral reef slopes below 30-40 m. Yet in spite
of their commoness in some areas, rarity or endangerment
in others, and their historical and commercial value, black
corals remain among the least known of colonial coel-
enterates. Almost nothing is known about the architectural
and material properties of black coral skeleton, which may
determine not only its commercial value, but its ecological
functions as well. This paper is the first attempt to elu-
cidate and compare the structural and mechanical prop-
erties of black coral skeleton, using as examples the geo-
graphically disparate congeners A. ftordensis and A. salix.
Materials and Methods
Colonies ofA.fiordensis were collected by SCUBA div-
ers at Doubtful Sound, New Zealand, in 10-20 m depth,
and A. salix at Cay Sal Bank, Bahamas, in 30-40 m. Liv-
195
196
K. KIM ET AL.
ing tissue was removed by jets of fresh water, and the
remaining skeletal material was stored dry at 4°C.
Microscopy
Skeletal material was examined both before and after
treatment with a variety of agents. Some skeleton was
examined after standard double fixation (3% glutaralde-
hyde and 1% osmium tetroxide) and double staining (sat-
urated uranyl acetate in 50% methanol and 0.3% lead
citrate). However, because the fibrils comprising the skel-
eton of both species are embedded in an amorphous, re-
sistant matrix, observation of fibrillar orientation or
structure often required chemical treatment. Unfixed
material was placed in concentrated formic acid for 24-
48 h at room temperature to cause swelling and delami-
nation of the skeletal layers (Goldberg, 1991). Using
watchmaker's forceps, narrow strips of the outer skeleton
were torn away along the long axis. Formic acid-treated
material was washed and dehydrated in graded ethanols
and embedded in Spurr resin. Thin sections were taken
at parallel, 45° and perpendicular to the long axis of the
branch. Additional material was examined after depro-
teination with 1 N KOH (24 h, 105°C; Hackman and
Goldberg, 1971). Sections were stained with 2% phos-
photungsticacid(PTA)in 10%ethanol (Bouligand, 1972)
and examined using a Phillips EM 300 transmission elec-
tron microscope (TEM) operated at 60 kV. Fine structural
observations of fibril orientation by scanning electron mi-
croscopy (SEM) could not be made without etching the
smooth surface of formic acid-treated material with 1 .0
M NaBH4 in 1 .0 M NaOH (4 h, 70°C). The delaminated,
etched skeleton was dried after ethanol dehydration and
coated with Au/Pd prior to examination in an ISI Super
3 A SEM. We also prepared fracture cross-sections for SEM
from acid-treated material. Samples of untreated, as well
as formic acid-treated, and KOH-treated skeleton were
examined in the light microscope. One micron sections
of Spurr-embedded material were either stained with To-
luidine blue, or viewed unstained using polarized light
and/or phase-contrast microscopy. We also examined
unembedded, aqueous mounts, and epoxy-embedded
thick sections ground and polished with graded abrasives.
Material properties
Measurements of density and Young's modulus were
performed on skeleton rehydrated in artificial seawater
(ASW). To estimate minimum time required for rehy-
dration, skeletal pieces of various diameters and lengths
were dried overnight at 70°C, and cooled in a desiccator.
The dry skeletons were weighed, placed in ASW (~37%o,
room temperature), then surface-dried and re-weighed
daily. When weights for three consecutive days were within
0.5%, hydration was considered complete and the gravi-
metric change was noted as the water content. The re-
hydrated piece was surface-dried and placed in a graduated
cylinder (accurate to 0.01 ml). Density was calculated as
weight of the skeleton divided by the volume of water
displaced.
Mechanical properties were determined using samples
6-10 cm long and 0.7-1.5 mm in diameter, embedded
in methyl methacrylate at both ends prior to rehydration.
We were limited to branches less than 1 .5 mm in diameter
because larger branches pulled out of the plastic before
failing. The plastic ends were loaded under tension using
an Instron 1011 strength tester fitted with a 5 kg cross
head travelling at a constant rate of 50 mm/min. The
Young's modulus (E) was calculated using the formula,
E = FL/AdL, where F and dL are the amount of force
applied to the branch and the change in its length at fail-
ure, respectively. The length of the skeleton (L) is the
distance between the two methyl methacrylate blocks,
while A is the cross sectional area of the sample. Because
the skeleton was never perfectly cylindrical and often ta-
pered toward the apex, an average cross sectional area
was estimated by assuming the taper was linear from one
end to the other. Hardness was measured on both the
Moh's and Vicker's scale. Moh's is a qualitative scale in
which 10 minerals are used as standards ranging from talc
(1) to diamond (10). Hardness is based on the ability of
one material to scratch another. The Vicker's scale is a
more quantitative measure of microhardness, which
measures the impression made using a pyramid-shaped
diamond forced into the surface of a material. Micro-
hardness is expressed as Vickers Hardness Numbers ac-
cording to the formula VHN == 1854P/d2 (kg mm"2),
where P is the load in grams and d is the mean length of
the indentation in microns (cf.. Hillerton el al., 1982).
Testing was performed both parallel and perpendicular
to the long axis of hydrated and unhydrated skeleton using
a Leco-DM 400F hardness tester with an applied load of
10 g.
Skeletal chemistry
Protein content and amino acid composition were de-
termined after hydrolysis in 4 N methanesulfonic acid
( 105°C, 20 h) (Simpson et al., 1976) using a JEOL 5AH
amino acid analyzer with a ninhydrin-based detection
system. Protein was expressed as total ninhydrin reactivity.
Chitin was estimated from the amino sugar content, cor-
rected for deacetylation. Dry skeletal powder was extracted
using chloroform:methanol (2:1. V: V) on a shaker for 48
h at room temperature. The powder was washed in meth-
anol (3X), dried overnight, and re-weighed to estimate
lipid content.
BLACK CORAL SKELETONS
197
Results
Microscopy
The external surfaces of antipatharian skeletons are
distinguished by the presence of spines. Both the spine
morphology and pattern of spination clearly distinguish
the two species, and correspond with skeletal characters
given by Grange (1988, 1990) and Opresko (1972). The
skeleton ofA.fiordensis is marked by rows of numerous,
smooth, slender spines 200-350 ^m long in addition to
shorter, branched secondaries (Fig. 1). In contrast, spines
in A. sa/ix are relatively uniform, compressed cones 90-
100 nm long, that become nodose during maturation (Fig.
5, inset). As is typical of antipatharians (Opresko, 1972),
the spines are organized into spiralling rows along the
long axis of the skeleton (Fig. 5). There are no secondary
spines in this species.
In light microscopic cross sections of A. fiordensis. the
growing tip can be distinguished clearly from more mature
portions of the skeleton, as a region that stains intensely
with Toluidine blue (Fig. 9). The skeleton increases in
thickness and in length by adding thin growth layers or
lamellae. These range in thickness from 0.1 to 1.0 ^m
when measured between the spines (Fig. 14). The lamellae
become thinner as they approach and add a layer to the
spines. Most of the layers are separated from each other
by a subtle discontinuity in optical density. However, at
irregular intervals, material that stains more intensely with
Toluidine blue separates sets of skeletal lamellae. These
deposits, which are also osmiophilic, are visually inter-
preted as growth rings. Growth ring structure has been
described by Goldberg (1991). Ring timing in this species
will be considered elsewhere (Grange and Goldberg, in
prep.).
Strongly birefringent patterns are observed when thin
pieces of formic acid-treated skeleton are examined with
polarized light microscopy. These patterns appear to result
from the chitin fibrils because subsequent deproteination
by KOFI, albeit incomplete (see skeletal chemistry section
below), does not appreciably affect the birefringence of
either species. Light microscopic observation of A. fior-
densis skeleton suggests that the fibrils form anticlockwise
helices around the long axis of the branch. Interpretation
of the longitudinal pattern is influenced by the presence
of numerous spines that serve as convergence points for
the fibrils within layers. The spines are numerous enough
to obscure the surface fibril pattern, making an accurate
assessment of its overall direction difficult. A view through
several lamellae creates the appearance of a meshwork
(Fig. 10), indicating a layer-to-layer change in the fibrillar
winding pattern.
Visualization ofA.fiordensis chitin microfibrils (here-
after referred to simply as fibrils) by transmission electron
microscopy requires treatment with formic acid, or formic
acid followed by KOH. In transverse section, the fibrils
constituting a lamella appear to be sub-parallel and woven.
At intervals, the fibrils are sparse, and it is here that the
lamellae separate with KOH or formic acid (Fig. 14, inset).
Although chemical treatment removes the electron-
opaque material that defines a lamella, it is clear that there
are several sublayers of fibrils within each of them. Sec-
tions of formic acid-treated material cut at 45° to the
skeletal long axis show tracts of fibrils within several ad-
jacent lamellae. Most of the tracts appear to intersect at
angles, giving them a twisted or cable-like appearance (Fig.
15). This pattern may be expected from helically wound
fibrils in successive layers that are out of phase. There are
no lamellae that clearly display fibrils with a preferred
orientation, and there are no differences in fibril pattern
when material is sectioned transversely. There are irreg-
ularly spaced regions of skeleton displaying parabolic or
arced patterns, suggesting a systematic, gradual rotation
of fibrils among adjacent lamellae (Fig. 15 and inset).
However, these are not as obvious in the electron micro-
scope as they are in the light microscope (see below).
Longitudinal preparations indicate that fibrils intersect
and overlap one another, displaying gradual changes in
orientation (Fig. 16). In some of our preparations there
are abrupt changes in orientation particularly in the spines,
where fibrils within layer are perpendicular to the long
axis. In addition, we note again that each lamella is com-
posed of multiple sublayers with different orientations.
Because some of these sublayers may only be 10 nm thick
(Figs. 14, inset; 15). thin sections of 90 nm may cut
through more than one set of fibrils. Thus the TEM pat-
tern that appears to show intersecting fibrils, may be an
artefact. In the SEM, the surface pattern of fibrils within
layer is clearer (Fig. 2). Spines are responsible for the large
changes in orientation at the top and bottom of the figure.
Figure 2 is a surface view of the multiple layers shown in
Figure 10. Fibrils appear to fan out between spines but
actually wind helically in the longer run.
If the etched surfaces of adjacent layers are examined
using SEM, large fibrillar biases (i.e., large angular devia-
tions) can be seen resulting from layer to layer changes
in fibril direction. In Figure 4, for example, 30 to 40°
changes occur in adjacent lamellae. Up to 45° changes
occur between layers in this species. The fracture pattern
also suggests a change in fibril orientation, resulting in
layers that fracture in different directions (Fig. 3). Perhaps
the most revealing view of this complexly arranged skel-
eton is the ground thick section shown in Figure 1 1 . This
transverse section shows that the skeleton is organized
into large-scale patterns that are difficult to see at the elec-
tron microscopic level. The cable-like arrangements of
fibrils appear to be a limited, thin-sectional view of mul-
198
K. KJM ET AL.
BLACK CORAL SKELETONS
199
tiple layers, each of which has small deviations in fibril
orientation. Constructive superimposition of fibrillar
tracts with similar orientation (but not parallel), give rise
to darker tracts of fibrils in polarized light. Regions be-
tween tracts give rise to lighter, diffuse areas, representing
regions of abrupt change in fibril orientation in successive
layers of skeleton. Thus, large (10-20 M). irregularly spaced
helicoids are formed that appear to rotate through 180°
(Fig. 15).
Light microscopic cross sections of A. sali.\ skeleton
are striking in their apparent uniformity compared to A.
fiordensis. This appearance is due to several factors. First,
there are no ontogenetic differences in the staining prop-
erties of the skeleton. Instead, clusters of lamellae alter-
nately stain ortho- and metachromatically with Toluidine
blue (Fig. 12). Second, the skeleton lacks the irregular
deposits of osmiophilic material seen as growth rings in
the New Zealand species (compare Figs. 1 2 and 9). Growth
rings in A. salix are very subtle and appear in unstained
material as slight but regular differences in spacing be-
tween clusters of lamellae. This subtle structural distinc-
tion all but disappears on staining with Toluidine blue.
Electron microscope observations suggest that rings cor-
respond to a particular arrangement of fibrils (see below).
This is a structural distinction in the formation of growth
rings in the two species. In A. fiordensis. rings are osmio-
philic and non-fibrillar (Goldberg, 1991). Polarized light
examination of KOH treated A. salix branches suggest
that the fibrils are wound around the long axis of the
skeleton following the gradual, left handed spiral pattern
of the spines. Fibrillar tracts between spines tend to be
more densely packed than they are around and within
the spines themselves (Fig. 1 3) in contrast to A. fiordensis.
Fibrils in successive layers reinforce this pattern resulting
in alternately opaque and translucent zones of birefrin-
gence parallel to the skeletal axis. This pattern suggests
the absence of large fibril biases as in A. fiordensis.
Transverse sections of A. salix skeleton (TEM) are or-
ganized into a series of light and dark bands that vary in
their specific characteristics, and range in thickness from
0.1 to 1.9 ^m. Individual lamellae appear to correspond
to these bands. The darkest regions consist of densely
packed fibrils arranged in parallel within a matrix with a
strong affinity for PTA. These grade into bands of inter-
mediate electron opacity that constitute the largest volume
of skeletal cross section. The intermediate bands may
contain fibrils in a number of planes, including parallel.
The lightest bands appear to be narrow (0.1-0.2 jtm) re-
gions where fibrils are obliquely arranged (Fig. 17). Al-
though it is difficult to determine with certainty, growth
rings appear to correspond to the juxtaposition of light
and dark bands within a series of lamellae (Fig. 18). Fibrils
in longitudinal sections occur in bands that remain parallel
for relatively long distances, and vary in their electron
opacity (Fig. 19). This structure corresponds to a section
through a band of intermediate electron opacity as seen
in cross section (e.g., upper part of Fig. 18), where fibrils
occur with minor variations in orientation.
Fracture preparations more clearly show the parallel
structure of the lamellae in A. sali.\ (Fig. 6). These result
from the smaller fibril biases between lamellae in this spe-
cies (compare with Fig. 3). In SEM preparations the sur-
face fibrils appear to be arranged in a chevron-like pattern
(Fig. 7). However, our overall view of the surface fibril
pattern in this species is limited due to its less tractable
response to chemical and physical manipulation. While
formic acid separates the skeleton into layers (Fig. 6, inset),
individual lamellae do not separate well, even with me-
chanical assistance. Additionally, because of the relative
uniformity of fibril distribution, phase contrast and po-
larized light microscopy is not useful in determining the
fibril pattern over long distances. Scanning electron mi-
croscopy is more successful in showing between-layer
changes in fibril orientation over short distances, although
etching with borohydride is not as effective in revealing
surface fibril patterns in this species. Changes in fibril ori-
entation do occur between layers, however, the fibril biases
are small (<20°) compared to A. fiordensis (compare Figs.
4 and 8).
Material properties
The Young's modulus is a measure of stiffness or ri-
gidity derived from a simple elastic behavior of materials
Figures 1-4. SEM preparations of Antipathes fiordensis.
Figure 1. Branch surface and its numerous and prominent spines; scale bar = 500 ^m.
Figure 2. Borohydride-etched branch surface showing complex pattern resulting from counterclockwise
winding of surface fibrils (arrows) around numerous spines (s). and the three-dimensional pattern of fibrils
in spines themselves. The spine at the top of the figure depicts the change in orientation of fibrils at its base;
scale bar = 10 pm.
Figure 3. Fracture pattern of formic acid-treated skeleton. 3b (scale bar = 5 |im) corresponds to the
boxed area in 3a (scale bar = 50 fim). Large fibril biases between layers result in lamellae that fracture in
different directions; arrows depict fibril orientation within the indicated layer.
Figure 4. Fibril biases of up to 45° (arrows) are encountered when comparing surface fibrils of successive
lamellae (formic acid-borohydride preparation); scale bar = 10 ^m.
200
K. KIM ET AL.
BLACK CORAL SKELETONS
201
in which stress is proportional to strain, although this pro-
portionality rarely occurs in most materials of biological
origin (Hepburn and Chandler, 1980). We calculated
Young's modulus from the sigmoidal load-deformation
curve of the antipatharian skeleton (e.g.. Fig. 20). A small
initial lag occurs as the specimen straightens under ten-
sion. This is followed by a linear region typically consti-
tuting >80% of the distance between the points of origin
and failure. The terminal plastic region, where the greatest
amount of deformation occurs just before failure, typically
constitutes less than 12% of the stress-strain relationship.
There is little deviation in the slope of the plastic region,
and failure almost always occurs at the test grips. Moduli
were calculated using only the linear portion of the curve,
and again, using the slope of a line drawn between the
origin and the point of failure. Only minor differences
were noted. However, we also found that using only the
linear portion was more prone to operator error. The
moduli reported here were determined by the second
method, which gives the average slope of the entire stress-
strain relationship. This method is therefore more con-
servative compared to those determined solely from the
slope of the more limited, linear portions of the load-
deformation curve (see Bassin el ai, 1979; Vincent and
Hillerton, 1979).
The material properties of the two antipatharians are
very different (Table I, Fig. 2 1 ). In addition to being stiffer
(more rigid), the skeleton of A. salix is more resistant to
deformation under tension than A. fiordensis (P < 0.000 1 ,
for both cases). Higher density (12%) is also characteristic
of A. sali.\. as is greater hardness. Hardness is a complex
measure of material strength and plasticity. Although there
are no absolute standards for hardness, a comparative
measurement adds an additional perspective for this little-
known material. The differences in skeletal hardness are
not resolved on the Moh's scale because both skeletons,
wet or dry, have hardnesses of 3, i.e.. both skeletons are
only hard enough to scratch calcite. However, calcite is
more readily scratched using A. salix. indicating that it is
slightly harder than A. fiordensis. Microhardness testing
further refines this observation, suggesting that A. sali.x is
approximately 17% harder and less variable along both
axes of testing compared to A. fiordensis (Table I).
Skeletal chemistry
We used amino acid analysis to estimate the relative
proportions of skeletal chitin and protein. Standard
methods of colorimetric analysis (e.g.. Lowry or Bradford)
consistently underestimate protein levels. Chitin can not
be estimated gravimetrically after "deproteination" with
KOH because a variable amount of protein remains as-
sociated with the chitin after treatment. Table II sum-
marizes these results. The skeletal tips are composed pri-
marily of protein and chitin. Extractable lipid is both low
and variable in the two species. The percent protein is
greater in A. fiordensis (P < 0.025) and, conversely, there
is a greater proportion of chitin in A. sali.\ (P < 0.05).
Overall, there are few differences in the amino acid com-
position of the skeletons (data not shown). When sequen-
tially treated with formic acid and aqueous borohydride,
some 6 to 9% of the protein is lost (Table III). Although
there are some differences in protein composition resulting
from this chemical treatment, the most significant change
is the total destruction of tryptophan. The dominant
amino acids, glycine, alanine, and histidine, are unaf-
fected, as are the levels of glucosamine (i.e.. chitin). The
principal effect of this reagent is the visual enhancement
of chitin fibrils in the scanning electron microscope (see
below).
Discussion
There are a number of commonalities between arthro-
pod cuticle and antipatharian skeleton. Some of the gen-
eral similarities in chemical composition have been de-
scribed recently (Holl et ai, 1992). Arthropod cuticle also
provides the basis of morphological comparison or con-
trast, because it is the most commonly studied example
of chitin-protein architecture. Like antipatharian skeleton,
cuticle is a composite material constructed of chitin fibrils
embedded in an amorphous protein matrix. Cuticle is
generally considered as a laminated structure with chitin
fibrils lying in parallel within each layer. Layer-to-layer
changes in fibril orientation have been characterized,
ranging from the rare, totally uniform fibril orientation,
to a helicoidal arrangement. Helicoids, as originally de-
scribed by Bouligand (1965), are optical artefacts of arced
Figures 5-8. SEM preparations of Antipathes salix.
Figure 5. Branch surface showing spiral pattern of spines: scale bar = 200 ^m. Inset: spines are laterally
compressed and nodose; scale bar = 50 /im.
Figure 6. Fracture pattern showing parallel orientations of adjacent lamellae; scale bar = 5 ^m. Inset:
formic acid treatment does not separate adjacent lamellae, but separates the skeleton into thicker layers,
possibly corresponding to growth rings; scale bar = 50 Mm.
Figure 7. Surface fibrils, revealed by borohydride-formic acid treatment, are more unidirectional compared
to A. fiordensis. and form convergent, chevron-like tracts; scale bar = 10 Mm-
Figure 8. Adjacent lamellae display relatively small fibril biases; mean direction is depicted by arrows;
treatment as in Figure 7; scale bar = 10 pm.
202
K. K.IM ET AL.
BLACK CORAL SKELETONS
203
or parabolic fibril patterns when multiple layers of parallel
fibrils are viewed obliquely or transversely. Helicoids are
formed by adjacent fibril layers that appear to rotate grad-
ually from parallel. There are several variations on this
theme. Growth layers, consisting of parallel chitin fibrils
formed during the day, alternate with helicoidal layers
deposited at night in some insect groups (Neville and
Luke, 1969). A "plywood" type of architecture can be
formed by some insects when parallel layers abruptly
change direction on each successive day by about 90°.
Alternatively, a pseudo-orthogonal "plywood" cuticle can
be formed by more gradual changes in orientation of daily
parallel fibril layers. In the latter case, helicoidal layers
may rotate gradually through 90° or 1 80° before a daily
parallel layer is deposited. Finally, some insects may not
deposit a daily parallel layer of chitin fibrils at all, thus
producing the appearance of continuous rotation (Neville,
1967, 1970; Barth, 1973). The helicoidal model is the
most common explanation of cuticular structure, es-
pecially among insects (cf. Filshie, 1982; Neville, 1984;
Hughes, 1987) and crustaceans (cf.. Bouligand, 1971;Gi-
raud-Guille, 1984; Compere and Goffinet, 1987). Chitin
helicoids have also been described from a whelk perios-
tracum (Hunt and Gates, 1984) and the test of a tunicate
(Gubb, 1975). While the helicoidal model has gained wide
acceptance, agreement is not universal. Important excep-
tions to the parallel fibril model include the presence of
vertical fibrils supporting the horizontal lamellae in some
arthropods (cf, Hepburn and Chandler. 1976). Others
have more generally disputed the helicoidal model, sug-
gesting that the layers of fibrils are indeed curved and are
not artefactual (Dennell, 1973; Dalingwater, 1975). He-
lically wound and crossed chitin fibrils occur in certain
insect groups but scant attention is paid to them in the
modern literature.
Both antipatharian skeletons are composed of helically
wound fibrils. They are unlike the crossed fibrillar chitin
described from insects, which form alternating layers of
left and right handed helices (Neville, 1967). Figure 22 is
a composite model of fibril structure, based on our mi-
croscopic observations. Antipathes salix skeleton is a
comparatively simple structure composed of layers with
small deviations in fibril orientation. Regions of parallel
fibril and sub-parallel orientation are common. Regions
of abrupt change in fibril orientation may constitute op-
tical discontinuities that appear as growth rings, but hel-
icoidally arranged layers as such, are absent. Antipathes
fiordensis skeleton is much more complex. Fibrils within
layer exhibit a more "active" pattern. There is a consid-
erable degree of angular change from one layer to the
next, as shown in both transverse and longitudinal sec-
tions. The helicoidal arrangement of the skeleton cannot
be depicted easily because the pattern is obscured by the
helically wound fibrils, the pattern of spination, and the
irregular thickness of skeletal layers. If these factors are
taken into account, the skeleton corresponds most closely
to a type D insect cuticle (Barth, 1973) characterized by
a helicoidal rotation of approximately 180° between par-
allel (in this case low angle, cable-like) fibrils.
A helical arrangement of the fibrils provides some flex-
ibility while preventing explosion and localized buckling
under multiaxial stress (Wainwright et al., 1976). How-
ever, in many helically wound structures there is a danger
of delamination caused by the incompatibility of strain
Figures 9-11. Light microscope preparations of Antipathes fiordensis.
Figure 9. Transverse 1-Mm thick section showing intensely stained (Toluidme blue) young skeleton,
followed by lightly stained region marked by irregular, dark growth rings. Spines at the periphery (arrowheads)
are obliquely sectioned; scale bar = 100 /im.
Figure 10. Formic acid-treated skeletal peel (unstained) examined with a combination of polarizing and
phase contrast optics shows the intersecting pattern of fibrils from several skeletal layers. Surface fibrils
converge around the bases of the spines (s); the long axis of the skeleton is parallel to the scale bar. Scale =
50 >/m.
Figure 11. Polished transverse section showing irregularly spaced, large scale helicoids. Cable-like fibrils
with acute angles constitute the helicoid boundaries (straight arrows). Apparent rotation of successive fibril
layers form the helix-like structures depicted by curved arrows. Fibrils converge with a longitudinally sectioned
spine at the bottom of the figure; scale bar = 50 fim.
Figures 12-13. Light microscopic preparations of Antipathes salix.
Figure 12. Transverse l-jjm thick section showing differential orthochromatic (dark layers) and meta-
chromatic (light layers) staining with Toluidine blue. The skeleton of this species is composed of multiple
lamellae, but does not have distinct, differentially staining growth rings; the incremental growth of spines
(arrowheads) is also shown; scale = 100 ^m.
Figure 13. Formic acid-treated skeletal peel (unstained) examined with a combination of polarizing and
phase contrast optics shows both convergent and parallel fibrils arranged in birefnngent light and dark bands.
In contrast to the New Zealand species, the fibrils tend to be more densely arranged between spines (s). This
view through multiple lamellae suggests that large fibril biases are not a prominent feature of this species;
the long axis of the skeleton is parallel to the scale bar; scale = 50 urn.
BLACK CORAL SKELETONS
205
—? Polnl ol Failure (xl.yl)
Average Slope= (y1-yO)/(x1-xO)
Origin (lO.yO)
468
Extension (%)
1 0
1 2
Figure 20. Typical load-deformation curve (solid line) from anti-
pathanan skeleton depicting the method of calculating Young's modulus
from the average slope (dashed line) of the stress-strain relationship.
Extensibility is the percent elongation of the specimen at failure.
between adjacent layers with different fibril orientations
(Wainwright et ai, 1976). One way of dealing with this
is to allow only small angular differences between fibrils
in adjacent layers as in certain types of helicoidally ar-
ranged insect cuticle. However, in antipatharian skeleton,
the layers are not simple laminated structures. Because
the spines are cemented and inserted layer upon layer,
the helically wound skeleton is fixed at multiple points.
We hypothesize that the spines increase the surface area
for cementing one skeletal layer to the next. Moreover,
they may play an important role as continuous rivets,
preventing delamination from shear forces produced by
skeletal bending and torsion. If this suggestion is correct,
the presence of spines should reduce or eliminate the re-
quirement of small fibril biases between helically wound
layers.
In addition to differences in fiber patterns, there are a
number of other disparities inherent in the diversity
among the insect cuticles that make mechanical compar-
isons with black coral skeletons difficult. There are male-
female differences, maturational differences, and regional
differences within single cuticles. There are also differences
in technique among investigators, some of whom have
apparently performed mechanical testing without taking
water content into account (see review by Vincent, 1 980).
The data given by Vincent show, not surprisingly, that
stiffness values in insect cuticle vary from 1 X 106 NirT2
to 17 GN irT2, with a mean of 1.8 GN m~2 for all 28
cases. While the Young's modulus of the black corals
studied here fall in line with the mean, the stiffest of insect
cuticle can be as stiff as compact bone (see Hepburn and
Joffe, 1976; Vincent and Hillerton. 1979). It is doubtful
that antipatharian skeleton from a comparable number
of species will be found with a comparable range of values.
In a composite structure, the fibrils can be expected to
stiffen the more deformable matrix by reinforcing it. The
degree of reinforcement should increase quickly with in-
creasing volume to about 10-20%, irrespective of the fibril
orientation, after which gains in stiffness become dispro-
Figures 14-16. Transmission electron microscope preparations of Anlipathes fiordensis. Scale bars =
1.0 /jm.
Figure 14. Transverse section of doubly fixed, doubly stained skeleton. Lamellae or growth layers are
defined by weak osmiophilia or subtle changes in electron opacity (between triangles). Growth rings are
perceived as darker or thicker regions of osmium deposition (triangles). Inset: treatment with formic acid
and KOH removes protein matrix and osmiophilic material (triangles). Lamellae are composed of fibrils in
a variety of orientations, and tend to separate where the fibrils may be diffuse or absent (transverse section).
Figure 15. Formic acid-treated, PTA-stamed skeleton sectioned at 45°: skeletal layers consist of light
and dark bands. The more prominent dark, cable-like fibrils are interpreted to result from the optical in-
tersection of helically wound fibril layers; light regions are areas where fibrils appear to undergo rotation.
Portions of larger scale helicoids. resulting from regular change in fibnl bias in successive lamellae, can be
seen by focusing on the overall pattern of apparent curvature (arrows). Inset: detail of a partial helicoid.
Figure 16. Formic acid-treated longitudinal section showing gradually intersecting (arrowheads) and
abrupt changes in fibril orientation (curved arrow). The latter can be due to the presence of spines (compare
with Fig. 2). or a section through more than one layer of fibrils; PTA stain.
Figures 17-19. Transmission electron microscope preparations of Anlipalhes sali\. Scale bars = 1.0
/im.
Figure 17. Transverse section showing lamellae with differing fibnllar orientations and electron opacities.
O = zone of obliquely oriented fibrils; P = region of parallel fibrils; and i = intermediate zone. Fibril
orientations parallel to the long axis of the skeleton are prominent in this species.
Figure 18. Detail of cross section showing electron opaque lamella with parallel orientation followed
by narrow, electron-lucent zone of obliquely oriented fibrils and zone of intermediate electron opacity with
several changes in fibril orientation; symbols as above.
Figure 19. Longitudinal section showing parallel and sub-parallel orientation (arrowheads) of fibnllar
tracts (compare with Fig. 7). All preparations doubly fixed; PTA stained.
206
K. KIM ET AL
Z
o
•a
O
5
a)
A, fiordensis
A. salii
Cross Sectional Area (10E-6 m2)
Figure 21. Scattergrams of (a) Young's modulus and (b) extensibility
of the skeleton in Antipathes fiordensis and .-1- salix plotted against cross-
sectional area. Young's modulus is determined from the average slope
of the load-deformation curve (see text); extensibility is the percent elon-
gation of the specimen at failure. The small degree of overlap in the data
reflects significant differences in mechanical properties. Statistics are given
in Table I.
portionately smaller (Wainwright et a/., 1976). Insect cu-
ticle is highly variable in its chitin content, ranging from
4 to 60% (Vincent, 1980). In stiff cuticle, the chitin content
tends to be 30 to 40% of the dry weight, while in more
pliant cuticle, the chitin content tends to be higher, on
the order of 50-60% of dry weight (Vincent, 1980; Hil-
lerton, 1984). Black coral skeleton examined thus far
contains a relatively low proportion of chitin, ranging from
6 to about 15% (Goldberg, 1978; this paper), and corre-
sponds to published Young's modulus values for stiff cu-
ticle. Although it is not yet possible to allocate mechanical
properties to specific components of the skeleton, the
greater chitin content of A. sali.x (~29% more than the
mean of A. fiordensis) is within the overall range of chitin
values where the difference may account for at least some
of the increased skeletal stiffness.
The antipatharians are less rigid than a number of other
biological materials, including wood, bone, mollusk shell,
and some insect cuticle, while having a density higher
than wood, lower than shell or bone, but about the same
as insect cuticle (Wainwright et ai, 1976; Table 5.3). The
ratio of Young's modulus to density (E/p) is the specific
modulus, a means of assessing the stiffness per unit weight
of materials. High values of specific stiffness are often
considered superior to low values because they enable
construction ofstifferand lighter structures. However, for
antipatharians, greater flexibility per unit of density should
be more important than stiffness. Thus antipatharians
have a lower specific modulus compared to insect cuticle
values given in Wainwright et a/., by having a lower mod-
ulus with about the same density (E/p = 5.1-7.9 for 2
insects, and 1.0-2.3 for .4. fiordensis and A. salix, respec-
tively). If the Young's modulus given for Cirripathes sp.
(E = 0.3 GN m~2) in Wainwright et al. (1976) is combined
with our density measurement of 1 . 1 in C. litetkeni Brook
from the Bahamas (see Goldberg, 1976 for description),
a skeleton of even lower specific modulus results (E/p
= 0.27). The Young's moduli of the two Antipathes species
differ by more than twofold, but range from 4.1 to 10.7
times stiffer than that reported for Cirripathes sp. by
Wainwright et al. (1976). Antipatharians of this genus,
unlike the ones we studied, are unbranched and whip-
Table l
Material properties of antipatharian skeletons
Properties
Anlipalhes fiordensis
Antipathes sali\
Test statistics
I) Young's Modulus (GN/m2)
1
24 (0.360);
n
= 31
3.20
(0.511);
n
= 23
X2
= 38.89;
P< 0.0001
II) Extensibility (mm/cm)
7,
37(1.667);
n
= 31
3.84
(0.783);
n
= 23
X2
= 36.95;
P< 0.0001
III) Density (g/cm3)
1
25 (0.096);
n
= 15
1.40
(0.058);
n
= 25
X2
= 19.45;
P< 0.0001
IV) Hardness (Mohs)
3
3
V) Microhardness(HV)
i) Long Axis
18.
2 (1.473);
n
= 3
22.1
(0.378);
n
7
X2
= 3.857;
P < 0.05
ii) Short Axis
20
3 (3.342);
n
= 3
22.8
(0.987);
n
= 3
X2
= 0.429;
P > 0.05
Means and standard deviations (in parentheses) are provided along with numbers of observations (n). The Kruskal-Wallis One-Way ANOVA by
ranks (Chi-squares corrected for ties) were calculated to note differences between the two species.
BLACK CORAL SKELETONS
207
Table II
Chemical composition ofantipatharian skeletal tips
Antipathes
Properties
fiordensis
Antipathes
salix
Test statistics
I) Water content (<!
b dry wt.)
21.1
(0.59); n
= 15
19.6
(0.72);
n
= 25
X2
= 20.63;
P< 0.0001
II) Chitin content (".
o org. wt.)
10.4
(2.20); n
= 8
14.7
(1.25);
n
= 3
X2
= 5.04;
P < 0.05
Ill) Protein content (
% org. wt.)
55.4
(2.93); n
= 8
51.3
(0.58);
n
•>
X2
= 6.00;
P < 0.025
IV) Lipid content (% dry wt.)
0.32
(0.70); n
= 25
0.22
(0.503);
n
= 25
X2
= 0.32;
P > 0.05
Means are followed by standard deviations (in parentheses) along with numbers of observations (n). Kruskal-Wallis One- Way ANOVA by ranks
(Chi-squares corrected for ties) were calculated to note differences between the two species.
like, often forming corkscrew-like shapes on cliff faces of
the reef. The low Young's modulus may structurally reflect
this idiosyncrasy. Unfortunately, the correlative architec-
tural properties of this genus are unknown.
Mechanical properties of gorgonian skeleton have been
reported (Goldberg et al. , 1 984; Jeyasuria and Lewis, 1 987;
Esford and Lewis, 1990), and although the two systems
differ structurally and chemically (collagen instead of chi-
tin, and distinct amino acid composition among other
differences) the Young's moduli of the two black corals
fall within the range (1.1-9.3 GN nT2) reported for the
Table III
Organic composition of skeletal powder after formic acid and sodium
boTohydride treatment compared to untreated materials
A nt ipat lies fiordensis
Antipathes salix
Amino acids
ASP
-6.8 (*)
5.9 (**)
THR
-9.6
0.8
SER
-2.4 (ns)
6.2 (*)
GLU
6.8
10.7
PRO
-14.7
9.5
GLY
2.6 (ns)
5.3 (*)
VAL
-13.7
-18.4
MET
-40.0
-10.0
ILE
-18.6
-11.7
LEU
0.1
4.3
TYR
-13.5 (ns)
-17.3 (*)
PHE
-39.1 (**)
-14.5 (ns)
HIS
8.1 (**)
7.6 (*)
TRP
-100.0 (**)
- 100.0 (*)
LYS
-56.2
-48.2
ARG
-10.6
0.1
Protein
-6.1 (ns)
-8.7 («)
Chitin
8.4 (ns)
13.7 (*)
(ns) not significant; (*) P < 0.05; (**) P < 0.025.
Changes are noted as percent decreases (negative values) or increases
(positive values) from organic composition of untreated skeleton. All
values are means of three trials and statistical differences were analyzed
as in previous tables.
tips of 13 gorgonians by Esford and Lewis (1990). Inter-
estingly, they found that stiffer axes were typical of species
from deeper water but unlike antipatharians, gorgonian
skeletons from such environments are often calcified.
There is a growing body of evidence showing a rela-
tionship between skeletal mechanics and ecological func-
tion. In organisms with flexible skeletons, orientation to
flow can maximize efficiency of suspension feeding and
minimize drag forces (reviewed by Wainwright et al.,
1976). In certain gorgonian corals, adaptation to flow may
be recorded in the skeleton as a change in preferred ori-
entation of fan-like species (Wainwright and Dillon, 1969;
Grigg, 1972; Velimirov, 1976). In branched gorgonians
the skeleton can be reinforced perpendicular to the di-
rection of flow, by deposition of carbonate (Wainwright
and Koehl, 1976; Wainwright et al, 1976). Preferred ori-
entation occurs in the Antipatharia (Warner, 1981), and
there is a degree of it exhibited in A. fiordensis. Colonies
near the mouths of the fiords, are subjected to more con-
sistent current fields, resulting in more fan-shaped colo-
nies. Otherwise, this species tends to branch in multiple
planes (Grange, 1988). In addition, the skeleton is often
elliptical in cross section, especially in the thicker
branches, with the compressed sides facing the predom-
inant current flow. In A. salix, there is no obvious struc-
tural asymmetry, and the colonies tend to be branched
in many planes. Antipatharians generally require low
Theological environments. Unlike other cnidarians, the
polyps have no structural protection from the abrasive
forces associated with strong current. The muscular sys-
tems of the polyps and tentacles are so poorly developed
that a modest contraction is their only apparent defense
against such forces (Goldberg and Taylor, 1989). Trans-
plant experiments into relatively shallow water further
suggest that abrasion is a major source of mortality (Grigg,
1965). Thus the substantial structural and mechanical
properties of the black coral skeleton seem to be overde-
signed for the deeper and hydrodynamically more docile
zones in which antipatharians are generally found. It
208
K. KIM ET AL.
A B
Figure 22. Composite sketch of fibnl patterns. (A) Antipathes xali.\ is shown with surface fibrils helically
wound in an anticlockwise direction. The gradual change in the rotational sense of the winding pattern is
shown in successive longitudinal sections. Spines are shown as the centers of the tibnllar pattern. Transverse
sections through adjacent layers depict gradual, angular changes in fibril orientation as well as layers with
little or no change. (B) Antipathes fiordensis is depicted with a swirling pattern of fibrils that generally tend
anticlockwise. The more numerous spines are shown as focal points for the surface fibnl pattern. Abrupt
layer-to-layer changes in fibnl onentation are characteristic of this species. No two adjacent layers have the
same fibril pattern.
seems counterintuitive to find such stiff skeletons in zones
of relatively low velocity water movement.
While the fit between ecological function and skeletal
design is unclear, the distinction between the two species
studied has shown that Antipathes salix is darker, harder,
more dense, more hydrophobic, and stiffer than A. fior-
densis. These material differences appear to reflect the
more considerable commercial value of A. salix in the
jewelry trade.
Acknowledgments
We thank W. Faulkner of Telectronics Pacemakers
Corp. for his assistance with microhardness testing and
R. Nutt for the illustrations. Both K. Gordon (Biology)
and C. Levy (Mechanical Engineering) at FIU contributed
substantially to the mechanical analysis. We also thank
K. Gordon, K. Grange, and two anonymous reviewers
for helpful comments on the manuscript. This work was
supported by NSF grant OCE-86 13884 (to W.G.). New
Zealand coral material was obtained with the assistance
of K. Grange and R. Singleton of DSIR, Wellington, and
the support of the U.S.-New Zealand Cooperative Science
Program, as well as funds from DSIR. The Florida Insti-
tute of Oceanography provided ship time and facilities to
support the collection of Bahamian coral material. Col-
lecting permits from the Governments of New Zealand
and the Bahamas are also gratefully acknowledged.
Literature Cited
Bassin, M., G. M. Brodsky, and H. Wolkoff. 1979. Statics and the
Strength of Materials, 3rd Ed. McGraw-Hill, Toronto.
Barth, F. G. 1973. Microfiber reinforcement of an arthropod cuticle.
Laminated composite matenal in biology. Z Zellforsch. 144: 409-
433.
BLACK CORAL SKELETONS
209
Bouligand, V. 1965. Sur une architecture torsadee repandue dans les
nombreuses cuticles d'arthropodes. C.R hcbd. Seances Acad. Sci.
Pans 261: 3665-3668.
Bouligand. V. 1971. Les orientations fibnllaires dans le squelette des
Arthropods 1 . L'example des crabes. 1'arrangement torsade des strates.
./ Miir<>sc. (Paris) 11:441-472.
Bouligand, Y. 1972. Twisted fibrous arrangements in biological ma-
terials and cholesteric metaphases. Tissue & Cell 4: 189-217.
Compere, P., and G. Goffinet. 1987. Aspects ultrastructureaux et fonc-
tionnels de diverses regions cuticulaires non mineralisees d'un crustace
decapode. Carcinus maenas. Ann. Soc. R Zoo/. Belg. 117: 159-173.
Dalingwater, J. E. 1975. The reality of arthropod cuticular laminae.
Cell Tissue Res 163: 411-413.
Dennell. R. 1973. The structure of the cuticle of the shore crab, Carcinus
maenas. Zoo/. J. Linn. Soc. 52: 159-163.
Esford, L. E., and J. C. Lewis. 1990. Stiffness of Caribbean gorgonians
(Coelenterata. Octocorallia) and Ca/Mg content of their axes. Mar.
Ecol. Prog Ser. 67: 189-200.
Filshie, B. K. 1982. Fine structure of the cuticle of insects and other
arthropods. Pp. 281-312 in Insect Ultrastructnre. Vol. 1. R. King
and H. Akai. eds. Plenum Press, New York.
Giraud-Guille, M. M. 1984. Fine structure of the chitin-protein system
in the crab cuticle. Tissue & Cell 16: 75-92.
Goldberg. W. M. 1976. A comparative study of the chemistry and
structure of gorgonian and antipatharian coral skeletons. Mar Biol.
35: 253-267.
Goldberg, \V. M. 1978. Chemical changes accompanying maturation
of the connective tissue skeletons of gorgonian and antipathanan
corals. Mar. Biol. 49: 203-210.
Goldberg, \V. M. 1991. Chemistry and structure of skeletal growth
rings in the black coral Anlipathes fiordensis (Cnidaria, Antipatharia).
Hydrobiologia 216: 403-409.
Goldberg, VV. M., J. Makemson, and S. B. Colley. 1984. Entocladia
endozoica sp. nov.. a pathogenic chlorophyte: structure, life history.
physiology and effect on its coral host. Biol. Bull 166: 368-383.
Goldberg. W. M., and G. T. Taylor. 1989. Cellular structure and ul-
trastructure of the black coral Antipalhes aperta: 1. Organization of
the tentacular epidermis and nervous system. / Morphol. 202: 239-
253.
Grange, K. R. 1985. Distribution, standing crop, population structure
and growth rates of black coral in the southern fiords of New Zealand.
NZJ. Mar Freshw. Res. 19: 467-475.
Grange, K. R. 1988. Redescription of Anlipathes aperta. Totton, (Coe-
lenterata: Antipathana), an ecological dominant in the southern fiords
of New Zealand. NZJ. Zoo/. 15: 55-61.
Grange, K. R. 1990. Anlipathes fiordensis, a new species of black coral
(Coelenterata: Antipatharia) from New Zealand. NZJ Zoo/. 17: 279-
282.
Grigg, R. 1965. Ecological studies of black coral in Hawaii. Pac. Sci.
19: 244-260.
Grigg. R. W. 1972. Orientation and growth in sea fans. Limnol.
Oceanogr 17: 185-192.
Gubb. D. 1975. A direct visualization of helicoidal architecture in Car-
cinas maenas and Halocynthia papillosa by scanning electron mi-
croscopy. Tissue & Cell 7: 19-32.
Hackman, R. H., and M. Goldberg. 1971. Studies on the hardening
and darkening of insect cuticles. J. Insect Physiol. 17: 335-347.
Hepburn, H. R., and H. D. Chandler. 1976. Material properties of
arthropod cuticles: the arthrodial membranes. J. Comp. Physio/. 109:
177-198.
Hepburn, H. R., and H. D. Chandler. 1980. Materials testing of ar-
thropod cuticle preparations. Pp. 1-44. in Cuticle Techniques in Ar-
thropods, T. A. Miller, ed. Springer- Verlag. New York.
Hepburn, H. R., and I. Joffe. 1976. On the material properties of insect
exoskeletons. Pp. 209-235 in The Insect Integument. H. R. Hepburn,
ed.. Elsevier. New York.
Hickson. S. J. 1924. An Introduction to the Study oj Recent Corals.
Manchester LIniversity Press, Longmans. Green & Co., London. 257
P
Hillerton, J. E. 1984. Cuticle: mechanical properties. Pp. 626-637 in
Biologv of the Integument 1. Invertebrates, J. Bereiter-Hahn, A. G.
Matolsky, and K. S. Richards, eds. Springer-Verlag, New York.
Hillerton, J. E., S. E. Reynolds, and J. F. V. Vincent. 1982. On the
indentation hardness of insect cuticle. J. Exp. Biol. 96: 45-52.
Holl, S. M., J. Schaefer, W. M. Goldberg, K. J. Kramer, T. D. Morgan,
and T. L. Hopkins. 1992. Comparison of black coral skeleton and
insect cuticle by a combination of carbon- 13 NMR and chemical
analyses. Arch. Biochem. Biophys. 292: 107-1 1 1.
Hughes, P. M. 1987. Insect cuticular growth layers seen under the
scanning electron microscope: a new display method. Tissue & Cell.
19: 705-712.
Hunt, S., and K. Gates. 1984. Chitin helicoids accompany protein hel-
icoids in the periostracum of a whelk, Biiccinum Tissue & Cell 16:
565-575.
Jeyasuria, P., and J. C. Lewis. 1987. Mechanical properties of the
axial skeletons on gorgonians. Coral Reefs 5: 213-219.
Neville, A. C. 1967. Chitin orientation in cuticle and its control. Adv.
Insect Physiol. 4:213-286.
Neville, A. C. 1970. Cuticle ultrastructure in relation to the whole insect.
Symp. R Entomol. Soc. Loud. 5: 1 7-39.
Neville, A. C. 1984. Cuticle: organization. Pp. 61 1-625 in Biology of
the Integument L Invertebrates. J. Bereiter-Hahn, A. G. Matolsky,
and K. S. Richards, eds. Springer-Verlag. New York.
Neville, A. C., and B. M. Luke. 1969. A two-system model for chitin-
protein complexes in insect cuticles. Tissue & Cell I: 689-707.
Opresko, D. M. 1972. Redescription and reevaluations of the anti-
patharians described by L. F. De Pourtales. Bull. Mar. Sci. 22: 950-
1017.
Simpson, R. J., M. R., Neuberger, and T. Y. Liu. 1976. Complete
amino analysis of protein from a single hydrolysate. J. Biol. Chem.
251: 1936-1940.
Velimirov, B. 1976. Variation in growth form ofEunicea cavolinii Koch
(Octocorallia) related to water movement. J. Exp. Mar. Biol. Ecol.
21: 109-117.
Vincent, J. F. V. 1980. Insect cuticle: a paradigm for natural composites.
Symp. Soc. Exp. Biol. 34: 183-210.
Vincent, J. F. V., and J. E. Hillerton. 1979. The tanning of insect
cuticle — a critical review and a revised mechanism. / Insect Physiol.
25: 653-658.
Wainwright, S. A., and J. R. Dillon. 1969. On the orientation of sea
fans (genus Gorgonia). Biol. Bull. 136: 130-139.
\\ainwright, S. A., and M. A. R. Koehl. 1976. The nature of flow and
the reaction of benthic Cnidaria to it. Pp. 5-2 1 in Coe/enterate Ecology-
and Behavior. G. O. Mackie, ed. Plenum Press, New York.
Wainwright, S. A., W. D. Biggs, J. D. Currey, and J. M. Gosline.
1976. Mechanical Design in Organisms. Edward Arnold Publ. Ltd.,
London. 423 pp.
Warner, G. F. 1981. Species descriptions and ecological observations
of black corals (Antipatharia) from Trinidad. Bull. Mar. Sci. 31: 147-
163.
Wood, E. M., and S. M. Wells, 1988. The Marine Curio Trade: Con-
servation Issues. A report for the Marine Conservation Society,
Herefordshire, England. Pp. 1-120.
Reference: Biol. Bull. 182: 210-220. (April, 1992)
Adult Plasticity and Rapid Larval Evolution
in a Recently Isolated Barnacle Population
PETER T. RAIMONDI
The Marine Science Institute and Department of Biological Sciences,
University of California, Santa Barbara, California 93106
Abstract. Balanus amphitrite, a common barnacle spe-
cies, was introduced into the landlocked Salton Sea in
1 943 or 1 944. In 1 949, Balanus amphitrite from the Salton
Sea was classified as the subspecies, Balanus amphitrite
saltonensis. based upon morphological differences be-
tween Salton Sea and coastal individuals. This classifi-
cation was maintained following an investigation of the
Balanus amphitrite complex, in 1975. Such a designation
implies that the morphological divergence is underlain by
genetic differences. Using field and laboratory transplan-
tations, I tested the alternative hypothesis that the ob-
served morphological divergence in the adult stage of
Balanus amphitrite was the result of phenotypic plasticity.
The results show that the divergence in the examined adult
characters is in fact due to environmentally induced phe-
notypic plasticity. There were also phenotypic differences
between larvae from the Salton Sea and those from coastal
habitats that only became apparent during experimen-
tation with the adult stage. Here, however, experimental
results suggest that the divergence was due to an evolu-
tionary process, probably selection. These results also
provide the basis for two slightly precautionary conclu-
sions: ( 1 ) the observation that individuals living in typical
and novel habitats differ cannot even weakly indicate a
cause for the difference, and (2) a consideration of the
divergence of populations is incomplete if all of the life
history stages of the organism are not studied.
Introduction
One of the continuing challenges in evolutionary ecol-
ogy is to determine the genetic contribution to phenotypic
variation among populations. Two general, and non-in-
dependent processes can cause such phenotypic differ-
Received 13 May 1991; accepted 6 January 1992.
entiation (Gould and Johnson, 1972; Berven etal., 1979;
Falconer, 1989); implicitly assumed in both cases is that,
within a species, all individuals share a common ancestor
and, therefore, are derived from the same ancestral ge-
notype. First, populations may evolve differently (evolu-
tion). Second, even when the gene frequencies of two
populations do not differ, phenotypic differences may re-
sult from plasticity in some traits (phenotypic plasticity;
see Smith-Gill, 1983; West-Eberhard, 1989 for a discus-
sion of the forms of phenotypic plasticity). The two pro-
cesses may also interact to produce phenotypic variation
among populations.
The distinction between phenotypic differentiation by
evolutionary mechanisms and differentiation resulting
from phenotypic plasticity cannot be made on the grounds
that the latter has a non-genetic basis (West-Eberhard,
1989). Phenotypically plastic responses (to the environ-
ment) have as much genetic basis as do other less plastic
characters, and plasticity is therefore a trait subject to
evolutionary change (Bradshaw, 1965; Williams, 1966;
Schlichting, 1986; Macdonald rta/., 1988; West-Eberhard,
1989). The distinction is simply that evolution is a char-
acteristic of populations, whereas plasticity is a charac-
teristic of individuals (after Lewontin, 1957). Thus, for
populations in which individuals exhibit no plasticity,
phenotypic modification as a response to the environment
is possible only at the level of the population (across gen-
erations). In contrast, for populations in which individuals
exhibit plastic characters, phenotypic modification in re-
sponse to the environment is possible at the level of the
individual (within a generation).
Either as untidy noise in complicated genetic systems
as it was once regarded (see West-Eberhard, 1989), or as
a selectable trait (Schlichting, 1986), plasticity is important
to measure. This is because without determining the con-
tribution of phenotypic plasticity, the adaptive significance
210
PLASTICITY AND EVOLUTION
211
of phenotypic variation cannot be assessed (Berven et al,
1979). In this study I investigated whether evolutionary
change or phenotypic plasticity was responsible for the
observed phenotypic variation between two Californian
populations of the barnacle, Balanus amphitrite. One
population was from a typical coastal (harbor) habitat in
San Diego, California, the other was recently isolated in
a novel environment, the Salton Sea.
The Salton Sea is a recently formed landlocked body
of saline water, the largest body of water in California. Its
average size is 55 by 24 kilometers, but the dimensions
vary considerably (Carpelan, 196 la). The genesis and re-
cent history of the Salton Sea have created an environment
that is in many ways different from an open marine en-
vironment (Carpelan, 1961b) and yet supports a simple,
but fascinating, community of introduced marine species.
In 1904-1905, a series of floods on the Colorado and Gila
rivers breached the headworks of an irrigation channel.
For two years these rivers, which normally drain into the
Gulf of California, emptied into the Salton Sink, a land-
locked sub-sea level basin in southeastern California, and
thus formed the Salton Sea. Since 1907, when the head-
works were repaired, the level of the Salton Sea has been
maintained by irrigation water, and its salinity has risen
from 3.65%o (Carpelan, 1961b) to about 43%o (Anony-
mous, 1989); the latter value is between 5 and 8%o greater
than is typical for ocean water. The visibility (usually less
than 1 m in the Salton Sea), ionic composition, chlorinity,
pH, dissolved oxygen, and temperature fluctuations ( 10-
36°C) of Salton Sea water differ from those of ocean water
(Carpelan, 1961b; Raimondi, pers. obs.); because it is an
inland body of water, tidal fluctuations in the Salton Sea
are trivial.
Balanus amphitrite was first noticed in the Salton Sea
in about 1943-1944. Apparently B. amphitrite individuals
were transported from the San Diego area by air during
Naval exercises, as adults on mooring buoys or ropes
(Cockerall, 1945; Newman and Abbott, 1980), or as larvae
in the bilge water of Naval flying boats (Hilton, 1945). By
late 1944 they were ubiquitous: ". . . they were already
multiplying so fast that a stick or board only had to be in
the water a few days before a crust of minute barnacles
started to form." (from Hilton, 1945). "Barnacles now
seem to outnumber all other forms of life, both vertebrate
and invertebrate, found in the Salton Sea." (from Cock-
erall, 1945).
In 1949, the Salton Sea population was described as
the subspecies Balanus amphitrite saltonesis (Rogers,
1949). Subspecific designation was supported by a mono-
graph by Henry and Mclaughlin (1975) on the barnacles
of the Balanus amphitrite complex; the authors distin-
guished between Balanus amphitrite amphitrite and Ba-
lanus amphitrite saltonensis on the basis of a multivariate
analysis of 15 morphological characters of individuals
taken from the field. However, Newman and Abbott
(1980) suggested that because the Salton Sea form was
also found in a population from Wilmington Harbor (Pa-
cific coast), the difference was ecotypic. Flowerdew (1985)
has recently recommended, based upon an electrophoretic
investigation of 3 1 alleles at 1 1 loci, that the subspecies
designation for the Salton Sea population be removed.
He found that the values from both indices of genetic
identity (I) and genetic distance (D) were in rhe range of
variation expected between conspecific populalions (Nei,
1972). He also concluded that there was no "significant
genetic differentiation" of Balanus amphitrite saltonensis
from Balanus amphitrite amphitrite. This implies that no
evolutionary divergence could have occurred between the
populations, which is incorrect. Nonsignificant I and D
values should only be viewed as not refuting the null hy-
pothesis that there is no divergence between populations
for the tested alleles (Richardson et al., 1986). Indeed,
there are cases of apparently separate (good) species
showing no electrophoretic divergence (Avise et al., 1975).
My initial interest was to determine whether the ob-
served morphological divergence between Salton Sea and
coastal adults (Henry and Mclaughlin, 1975) was due to
environmental factors. As the samples used in this initial
study came from field collections, there was no way to
determine the contribution of the environment to the di-
vergence. In the present study I made no attempt to eval-
uate any adult character other than those that had been
described as differing between the two populations. This
was because: (1)1 was interested in determining the basis
for the differentiating characters, and (2) the selection of
additional characters would have been largely unmoti-
vated, because, unlike larvae (see following), when adults
from different populations were reared under experimen-
tal conditions, they could not be distinguished.
While evaluating the mechanisms determining adult
morphological divergence, I found a number of differences
in the larvae of the two populations. The basis of these
differences was also examined.
Materials and Methods
Study organism, sites, and general methods
Balanus amphitrite is a moderately sized bay barnacle
[average basal diameter is between 15 and 20 mm (New-
man and Abbott, 1980)], with a virtually world-wide dis-
tribution (Henry and Mclaughlin, 1975). Like most
thoracican barnacles, it is a simultaneous hermaphrodite
(Strathmann. 1987), and fertilized eggs are brooded in the
mantle cavity of the parent until they become at least
stage one nauplius larvae, when they are expelled into the
water. In acorn barnacles such as Balanus amphitrite,
there are typically seven larval stages (Strathmann, 1987):
six naupliar stages (feeding) followed by the final cyprid
212
P. T. RAIMONDI
stage (non-feeding). All stages are potentially planktonic,
but stage one nauplii will often stay within the mantle
cavity of the parent, making stage two nauplii the first
planktonic stage (Raimondi, pers. obs.).
Adults were collected from three locations: ( 1 ) Salton
Sea, near North Shore, (2) Mission Bay, California, and
(3) Beaufort, North Carolina (larvae from these individ-
uals represented a second coastal population). After col-
lection, individuals were maintained in the laboratory at
a water temperature of 20-23°C, and were fed a mixed
diet of brine shrimp and the diatom, Skeletonium costa-
tum (see Rittschof et al, 1984). As individuals died, new
ones were brought in from the field so that 300-500 adults
per population were maintained at all times. Adults were
induced to expel brooded larvae by a combination of
overfeeding and direct bright light. Expelled larvae could
then be attracted by a light source and collected.
Larvae were grown in culture at 27-28 °C on a diet of
Skeletonium costatum (see Rittschof?/ al., 1984, for de-
tails of culturing techniques). Larvae from each population
were grown in separate containers (usually 3000-5000
per population in 10 1 of seawater). The larvae from each
of these rearing events were called a batch. Usually,
batches of larvae from all populations were reared si-
multaneously. With this protocol, individuals could not
be considered replicates for among-population compari-
sons, because the effect of batches could not be separated
from the effect of populations. Hence, for the examined
larval characters, the average value for the individuals
within each batch was used as the replicate unit.
All of the larvae used in the experiments described be-
low were reared at the Duke University Marine Labora-
tory in Beaufort, North Carolina. Upon metamorphosis
to the cyprid stage, individuals were collected and shipped
live in cold packs, via overnight delivery, to the University
of California at Santa Barbara.
Adult characters
As stated, Henry and Mclaughlin (1975) compared the
Salton Sea population with coastal populations using a
multivariate analysis of 1 5 morphological characters. The
statistical difference between populations was largely due
to six ratios of four measurements of the tergum (Table
I, Fig. 1 ). To determine the contribution of environmental
factors to the morphological divergence, as manifested in
these ratios, I did the following experiment. [The best
method of determining whether the morphological di-
vergence between populations was due to environmental
differences would have been to reciprocally transplant
newly settled individuals from one location to the other
(Mission Bay to Salton Sea, and vice versa). Legally and
ethically this could not be done].
Cyprids from both the Salton Sea and Mission Bay
brood stocks (see above) were allowed to settle on 10 X 10
cm clay tiles in the laboratory and raised to maturity on
those tiles in two environments: "lab," and "lagoon" (the
rationale for having two experimental habitats is given
below). The density of settlers was about 1 cm"2. Lab
individuals were grown under laboratory conditions in
the running unfiltered seawater system at the University
of California at Santa Barbara. Water temperature during
the experiment was about 20°C. Lagoon individuals were
raised at the same time as the lab individuals shallow salt
water lagoon (approximately 10 hectacres in size) on the
campus of the University of California at Santa Barbara,
however the water temperature in the lagoon during the
experiment varied between 25 and 28°C. The lagoon is
separated from the ocean by a sandy barrier through which
water, but not plankton, can pass. Salinity in both envi-
ronments was 32-33% during the experiment. No spon-
taneous (additional) settlement otBalanus ampfiitrite oc-
curred in either the lab or lagoon.
Lab individuals were fed a mixture of brine shrimp and
Skeletonium (see above); lagoon individuals fed upon the
natural plankton in the lagoon. When the lab and lagoon
individuals had grown to 6 to 8 mm basal diameter, they
were collected. Individuals of the same size were also col-
lected from both the Salton Sea and Mission Bay; these
were the "field populations" in all comparisons. In sum-
mary, there were six populations of barnacles: Mission
Bay — field, lab, and lagoon; and Salton Sea — field, lab,
and lagoon.
From the several hundred individuals reared or col-
lected from each population, 19-54 were randomly and
sequentially selected, and from each the tergum was re-
moved and placed individually in a small container of
bleach. This procedure removed all tissue from the cal-
careous mass. Differences in sample size reflect differences
among populations in the variability associated with
measurements (see Sokal and Rohlf, 1981). Each tergum
was drawn using dissecting microscopic and camera lucida
projection. From the drawings, measurements of the four
tergal dimensions were made and tergal ratios were cal-
culated (Table la). Ratios were compared among popu-
lations using a multivariate analysis of variance. Ratios
were used so the analysis would be comparable to that
done in the original work by Henry and Mclaughlin
(1975), which described the morphological divergence.
However, there are convincing arguments that the use of
ratios in morphometric analyses might lead to spurious
interpretation of data (Atchley et al., 1976). For this rea-
son, I also compared populations using the four characters
(not the ratios. Table Ib) in a multivariate analysis of co-
variance, as advocated by Atchley et al. (1976).
Two experimental habitats were tested because plastic-
ity can be a heritable trait, and the degree of expressible
plasticity, if any, might therefore have differed between
populations. At the extreme, one population might be
PLASTICITY AND EVOLUTION
213
plastic for the examined characters and the other might
not be. With only one experimental habitat, there would
have been no a priori way to control for this possibility.
In the following discussion I assume that phenotypic plas-
ticity, if any, will be in the form of phenotypic modulation
(Smith-Gill, 1983); this is a reasonable assumption for
characters like the ones examined (Table I). Suppose that,
in addition to the field populations, there were only the
lab populations, and that a posteriori analyses indicated:
( 1 ) no difference in the examined characters between the
Salton Sea and Mission Bay lab populations, (2) no dif-
ference between the lab populations and the Mission Bay
field population, and (3) that the lab and Mission Bay
field populations were all different from the Salton Sea
field population. Under this scenario there would be no
basis to support the hypothesis that the Salton Sea pop-
ulation was plastic for the examined characters but the
Mission Bay population was not, over the alternative hy-
pothesis that both populations were plastic and that lab
conditions are similar to conditions in Mission Bay. With
the inclusion of a second experimental habitat in the de-
sign, the former hypothesis could be ruled out if the Mis-
sion Bay lagoon population differed (in the examined
characters) from the Mission Bay lab and field popula-
tions. If such differences were not observed, then plasticity
in the examined characters would not be supported for
in the Mission Bay population. A problem could arise if
lab, lagoon, and Mission Bay habitats were all similar in
the characteristic that induced the plastic response. The
test of this would be the comparison of Salton Sea lab
and lagoon populations. As the Salton Sea population, in
this hypothetical case, was already shown to be plastic, if
the examined characters did not differ between the two
experimental populations it would suggest that the two
habitats were similar in a critical way. Other possibilities
concerning the degree of plasticity between populations
could be addressed following similar logical steps.
Larval characters
Three characteristics of the cultured larvae differed be-
tween the Salton Sea and Mission Bay populations: (1)
cyprid pigmentation — Salton Sea cyprids were unpig-
mented and white, whereas Mission Bay cyprids were
greenish-brown, (2) cyprid length — Salton Sea cyprids
were larger than those from Mission Bay, and (3) duration
in naupliar stages — individuals from the Salton Sea took
longer to become cyprids than did individuals from Mis-
sion Bay. I did the experiment described below to assess
the contribution of environmental factors to the diver-
gence in these larval characters and to correct for a lim-
itation of my initial observations: I did not know if Mis-
sion Bay larvae were representative of coastal larvae in
general. Although Henry and Mclaughlin (1975) surveyed
adult characteristics for a number of Balanus amphitrite
populations and found that coastal individuals were sim-
ilar, there have been no comparisons of larval character-
istics across populations for this species. Hence, I knew
that Mission Bay adults were representative of typical
coastal populations for the examined characters, but I had
no idea about the scale of phenotypic differentiation
among populations of larvae.
I compared cyprid length and pigmentation (color), and
duration in the naupliar larval stages, of individuals from
the Salton Sea, Mission Bay, and Beaufort, North Caro-
lina, that were reared in a laboratory under identical en-
vironmental conditions. Beaufort larvae were compared
to ones from Mission Bay to determine the extent of di-
vergence between geographically well-separated coastal
populations (i.e.. Atlantic vs. Pacific populations). Two
categories of larvae were used in experimentation: Gl and
G2. Gl larvae were progeny of adults brought from the
field to the lab and used as brood stock. To minimize the
effect of the parental environment, the first release of lar-
vae, which may have developed within the brooding adults
in the field, was not used. Some of the Gl larvae were
raised to maturity under laboratory conditions and their
progeny, G2 larvae, were also examined as a further con-
trol of residual parental effects. No G2 Beaufort larvae
were cultured because comparisons of Mission Bay Gl
and Beaufort Gl larvae indicated that these two coastal
populations did not differ for the examined larval char-
acters.
There are two general methods for defining and mea-
suring color (from Chamberlin and Chamberlin, 1980):
( 1 ) visual comparison with a standard that is accepted as
a reference, and (2) instrumental measurement of the
fundamental make-up of the constituent parts of the color
in terms of the relative contribution of absorption and
reflectance of each wavelength. Both methods were used.
For each batch of cyprids, 2000-3000 from each pop-
ulation were put in separate test tubes (cyprids from each
population in one test tube) and chilled to 6°C. This pro-
cedure did not damage the larvae, and it caused them to
congregate in the bottom of the tubes. The color of the
mass of cyprids was then compared to standards contained
in the Met/men Handbook of Colour (Kornerup and
Wanscher, 1978). No statistics are possible for this type
of color definition, therefore the Methuen coding will be
reported for reference.
To quantify an aspect of coloration, microspectropho-
tometry was performed on two batches of each larval
population. In initial sampling I found that there was di-
vergence in light transmittance between Salton Sea and
coastal populations in the range of 450 to 700 nanometers.
For logistical reasons I decided to concentrate compari-
sons on a particular wavelength and chose 510 nano-
meters. Transmittance was measured through a 40 X 40
214
P. T. RAIMONDI
Table la
Morphological characters used in the muhivariate analysis of variance
(MANOVA; explanation for the two tests is found in the text).
See Figure 1
1 ) The width of the tergal spur (sw)/the length of the basal margin
(bm).
2) The distance from the basiscutal angle to the margin of the spur
(aw)/the length of the tergal spur (si).
3) sl/sw.
4) aw/sw.
5) aw/bm.
6) sl/bm.
jim section in the middle of each cyprid. Light intensity
was standardized prior to each measurement.
Cyprid length was measured with a compound micro-
scope and micrometer. The final larval character that was
examined, the rate of larval development, required indi-
viduals to be drawn from culture and viewed microscop-
ically. This process can damage larvae and potentially
can introduce bacteria or ciliates to the culture. To min-
imize the risk of larval damage or culture contamination,
cultures were checked only once each day to determine
the developmental stage of the larvae.
Results
Adult characters
Tergal plate ratios or dimensions (Table la-b. Fig. 1)
for the six populations were compared in MANOVA and
MANCOVA procedures and there was a significant dif-
ference between populations (Table II). There were no
qualitative differences between the results of the two anal-
yses (MANOVA, MANCOVA), indicating that the use of
ratios would not, for this data set, lead to spurious inter-
pretations. Comparisons among populations clearly
showed where the differences were (Table II). The field
populations (Salton Sea vs. Mission Bay) were different
from each other, as also shown by Henry and Mclaughlin
(1975), and were different from all other populations.
However, when grown under similar conditions, there was
no difference between Salton Sea and Mission Bay indi-
viduals: Mission Bay and Salton Sea lab populations were
not significantly different, nor were Mission Bay and Sal-
ton Sea lagoon populations. Also, the two lab populations
(pooled for comparison) were different from the two la-
goon populations (also pooled). Examples of the plates
can be seen in Figure 2. Particular attention should be
directed to the tergal spur (see Fig. 1 for a detailed diagram
of the tergum). These results indicate that the phenotypic
differences between field populations in the Salton Sea
and Mission Bay are the result of phenotypic plasticity
and not genetic divergence.
table Ib
Morphological characters used in the muhivariate analysis of
covariance (MANCOVA). Basal margin (bm) was used as the
covariate. See Figure 1
1 ) The width of the tergal spur (sw).
2) The distance from the basiscutal angle to the margin of the spur
(aw).
3) The length of the tergal spur (si).
Plasticity itself is a trait that can be selected (Schmal-
hausen, 1949; Bradshaw, 1965; Schlichting, 1986), and it
could be argued that individuals from one of the two lo-
cations (Salton Sea and Mission Bay): (1) might not be
plastic, or (2) might not be as plastic as individuals from
the other location (Schlichting, 1986). If individuals from
one location were not plastic for the examined characters,
then there would be no statistical difference between lab,
lagoon, and field populations. For individuals from both
locations, there were highly significant differences among
all experimental populations (Table II). Thus, there is no
doubt that individuals from both locations are phenotyp-
ically plastic. No conclusive answer may be given to the
question of whether one population is more plastic than
the other because the degree of plasticity in individuals
from the two locations was not directly examined. How-
ever, the data suggest that individuals from the two lo-
cations are similar in their plasticity (in the examined
characters) because there were no differences between
Figure 1. The morphological measurements made on tergal plates:
( 1 ) The length of the basal margin. (2) the length of the tergal spur, (3)
the width of the tergal spur, (4) the distance from the basiscutal angle to
the margin of the spur (see Table I for examined ratios).
PLASTICITY AND EVOLUTION 215
Table II
Multivariate (MANO\'A & MANCOl'A) comparisons of tergal plate measurements between six populations <>/~Balanus amphitrite
(Table la-b. Fig. 1)
Pillai trace statistic
MANOVA
DF
F-STAT
P-VALUE
1.035
Pillai trace statistic
30. 790
6.878
<0.0001
MANCOVA
DF
F-STAT
P-VALUE
0.818
Population comparisons
15,474
11.846
<0.0001
COMPARISONS
MANOVA
P-VALUE
MANCOVA
P-VALUE
Conclusion
1) Mission Bay field vs. Salton Sea field
2) Mission Bay lab vs. Salton Sea lab
3) Mission Bay lagoon vs Salton Sea lagoon
4) Mission Bay field vs. both lab populations
5) Salton Sea field vs. both lab populations
6) Mission Bay field vs. both lagoon populations
7) Salton Sea field vs. both lagoon populations
8) Both lab populations v.v. both lagoon populations
<0.0001
0.594
0.340
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
0.869
0.247
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
Populations differ
No difference between populations
No difference between populations
Populations differ
Populations differ
Populations differ
Populations differ
Populations differ
Populations: Mission Bay field (n = 26). Salton Sea field (n = 45), Mission Bay lab (n = 28). Salton Sea lab(n = 26), Mission Bay lagoon (n = 21),
Salton Sea lagoon (n = 19). both lab populations (pooled n = 54), both lagoon populations (pooled n = 40). Because eight comparisons were made
(for each model) the critical P-VALUE for the population comparisons should be 0.05/8 = 0.0063.
reared Salton Sea and Mission Bay individuals in both
experimental habitats.
Larval characters
For each of the measured parameters, larvae from the
Salton Sea differed from the other two populations, which
were similar. Differences in pigmentation can result from
differences in food type, however, in these experiments
food type was constant among populations of parent
stock and larvae. The most noticeable difference between
Salton Sea cyprids and coastal ones was the lack of pig-
mentation in the former. Coastal cyprids were consis-
tently green-brown, whereas those from the Salton Sea
were white [Number of batches: Salton Sea Gl (8), Salton
Sea G2 (2), Mission Bay Gl (7), Mission Bay G2 (2),
Beaufort Gl (5)]. As compared to the Methuen color
standards (Kornerup and Wanscher, 1978), the color of
individuals from the Salton Sea was white (standard
3A1), while that for individuals from either of the coastal
populations was olive (standards 3F8-3F4). Individuals
from the two coastal populations were indistinguishable
on the basis of color. The results from a microspectro-
photometric analysis at 510 nanometers substantiated
the finding that pigmentation differed between individ-
uals from the Salton Sea and coastal populations (Table
III, Fig. 3).
Cyprid length also differed between Salton Sea and
coastal populations, which were similar (Table IV, Fig.
4). The third measured parameter was the time between
release of larvae by an adult and the metamorphosis from
the 6th naupliar larval stage to the cyprid stage (Fig. 5).
No analysis was performed on these data as there was no
way to meet an often unrecognized shared assumption of
parametric and nonparametric statistics: similarity of dis-
tributions among groups (Day and Quinn, 1989). Naupliar
duration was invariant among all populations except Sal-
ton Sea G 1 , and therefore there is no way to homogenize
variance terms. However, it should be obvious without a
probability value that naupliar duration was longer for
the Salton Sea populations than for the coastal popula-
tions.
In all cases where it was examined, within a population
there was no statistical difference between Gl and G2
cyprids indicating that residual environmental effects did
not affect the results (Figs. 3-5).
216
P. T. RAIMONDI
Figure 2. Tergal plates: (Top) Field populations: Mission Bay (left).
Salton Sea (right). (Middle) Lab populations: Mission Bay (left). Salton
Sea (right). (Bottom) Lagoon populations: Mission Bay (left). Salton Sea
(right).
Discussion
The linkage of evolutionary arguments to ecological
observations has been rather severely criticized in recent
Table III
A comparison of transmittance of light at 510 nanometers through
cyprids from five populations: (1) Salton Sea Gl. (2) Salton Sea G2,
(3) Mission Bay Gl. (4) Mission Bay G2, and (5) Beaufort Gl
ANOVA
Source
df
MS
F
P
Population
4
38.65
15.28
0.0052
Residual
5
2.53
For all populations, two batches of cyprids were examined. A-posteriori
comparisons are shown in Figure 3.
years. These criticisms are in two forms. First, the assign-
ment of specific evolutionary mechanisms to phenotypic
divergence has been questioned on the logical grounds
that most investigators postulating such mechanisms did
not properly test alternative hypotheses (Connell, 1980;
Underwood, 1990; but see Roughgarden, 1983). The sec-
ond criticism has been directed at investigators who failed
to consider genetic constraints when proposing evolu-
tionary explanations for ecological data (Gould and
Lewontin, 1979: Lande, 1979, 1982; Templeton. 1981;
Lynch, 1984). For an examination of phenotypic diver-
gence of an isolated population in a novel environment,
like Balanm amp/iilritc in the Salton Sea, understanding
these criticisms is crucial because phenotypic modification
of individuals in the novel environment is the expected
result of either evolutionary or plastic processes (see End-
ler, 1986). Hence, the observation that individuals differ
between coastal habitats and the Salton Sea cannot even
CYPRID PIGMENTATION
O
50
45
40
35
LH G2
i r
Salton Sea Mission Bay North Carolina
POPULATION
Figure 3. Transmittance of visible light, 510 nm, through cyprids of
three populations (no G2 Beaufort cyprids were cultured). Groups not
connected by horizontal lines differ at P < 0.05 [ANOVA with Tukey
procedure, see Table III]. Error bars are ± one standard error of the
mean.
PLASTICITY AND EVOLUTION
217
Table IV
A comparison ofcypricl lengths from five puinilalions: (I) Sa/ton Sea
Gl (S batches). (2) Salton Sea G2 (2 batches). (3) Mission Bay Gl
(7 batches). (4) Mission Bay G2 (2 bathces). ami (5) Beaufort Gl
(5 batches)
CYPRID LENGTH
ANOVA
Source
df
MS
F
P
Population
4
1954.16
20.25
<0.001
Residual
19
96.49
A-posteriori comparisons are shown in Figure 4.
weakly indicate a cause for the difference. The underlying
causes for phenotypic divergence in such populations can
be determined only through properly designed experi-
ments.
The first goal of this investigation was to determine
experimentally if the observed morphological divergence
between adult Balanus amphitrite in the Salton Sea and
those in coastal populations was due to evolutionary or
plastic processes. Like Henry and Mclaughlin (1975), I
too found that field populations of adult Balanus differed
for a number of characteristics. However, these differences
disappeared when individuals from the two locations were
reared in similar environmental conditions (laboratory or
lagoon). This is unequivocal evidence that the divergence
in the examined characters was due to phenotypic plas-
ticity.
During the investigation of adult characteristics, I found
that larvae from the Salton Sea differed from those from
Mission Bay. In subsequent experiments, I also found that
Salton Sea larvae differed from ones from another coastal
population, Beaufort, North Carolina, and that individuals
from the two coastal populations did not differ in any
examined larval character. The latter result is important
because it indicates that widely separated but coastal pop-
ulations have not diverged for the examined characters.
However, it should be noted that coastal populations of
bay or harbor species like Balanus amphitrite are probably
never completely isolated because of transport of adults
and larvae by ships (Carlton, 1985). The phenotypic dif-
ferences between the Salton Sea and coastal populations
persisted, undiluted, after two generations in the labora-
tory, suggesting that the differences are underlain by ge-
netic variation. Genetic crosses are needed to confirm this
suggestion (Falconer, 1989), however, in vivo crosses
would have been confounded by the possibility of self-
fertilization (Patel and Crisp, 1961), and ;'/; vitro crosses
that were attempted were unsuccessful.
Assuming that there is a genetic basis for the phenotypic
differences found between Salton Sea and coastal larvae,
what mechanism may be responsible for the divergence?
Only two mechanisms seem plausible: selection and ge-
500
OJ
Cfl
CO
CC 475
U
450
425
EHG2
N/A
Salton Sea Mission Bay North Carolina
POPULATION
Figure 4. Lengths of cyprids of three populations (no G2 Beaufort
cyprids were cultured). Groups not connected by horizontal lines differ
at P < 0.05 [ANOVA with Tukey procedure, see Table IV]. Error bars
are + one standard error of the mean.
netic drift, and of these I contend that selection is more
likely because there is evidence that there has been no
genetic drift. If genetic drift were responsible for the di-
vergence in larval characters for individuals in the Salton
Sea then: ( 1 ) the Salton Sea population must be isolated
from coastal populations, (2) the genes coding for the
characters that have diverged must be subject to very little
selection (stablizing selection), and (3) the effective pop-
ulation size of the Salton Sea population must have at
some time been small (after Falconer, 1989).
If these conditions were all met. then evolution by ge-
netic drift would probably occur. This would likely be
NAUPLIAR DURATION
5.5
Salton Sea Mission Bay North Carolina
POPULATION
Figure 5. Time between release of larvae from adults and the meta-
morphosis from the 6th naupliar stage to the cyprid stage (no G2 Beaufort
cyprids were cultured). Number of batches: Salton Sea Gl (8), Salton
Sea G2 (2). Mission Bay Gl (7), Mission Bay G2 (2), and Beaufort Gl
(5). Error bars are ± one standard error of the mean.
218
P. T. RAIMONDI
reflected in an electrophoretic comparison of allozymes
between the Salton Sea and conspecific populations be-
cause many of the genes coding for these enzymes would
probably be (effectively) selectively neutral (Falconer,
1989). In such a comparison, Flowerdew (1985) found
no evidence for either divergence in allelic proportions
or loss of heterozygosity for the Salton Sea population of
Balamts amphitrite. This indicates that significant genetic
drift has not occurred in the Salton Sea population. The
most likely reason that drift has not occurred is that the
inoculation population of the barnacle was not small
enough to promote a significant loss of heterozygosity
[heterozygosity is lost at a rate of 1/2 Ne per generation,
where Ne = effective population size (Lande, 1980)], and
that after the introduction its size increased explosively
(Cockerall, 1945; Hilton, 1945).
What selective agents could have caused the divergence
of larval characters in Salton Sea Balanus amphitrite? As
mentioned in the introduction, water in the Salton Sea
differs from that in typical oceanic habitats in a number
of ways. One that may be important in the present dis-
cussion is clarity. Ultraviolet (UV) radiation is harmful
to many marine organisms (Jokiel, 1980), and there is a
positive relationship between penetration by ultraviolet
radiation and water clarity (Jerlov, 1950). Pigmentation
has been proposed as an adaptive defense in marine or-
ganisms against damage by solar ultraviolet radiation
(Ireland and Scheuer, 1979; Yentsch and Yentsch, 1982;
Dunlop et ai, 1986). I suggest that pigmentation may
have been lost by cyprids in the Salton Sea because, in
part, the potential for damage by UV radiation in its
chronically turbid water is much lower than in coastal
water.
The other two characters showing divergence were cy-
prid length and naupliar duration (the series of naupliar
stages constituting the planktonic period prior to meta-
morphosis to the cyprid stage). Naupliar duration was
longer and the resulting cyprids were larger for individuals
from the Salton Sea than for those from coastal popula-
tions. The increase in cyprid size probably is, at least in
part, due to the increase in naupliar period, the period
during which larvae feed and grow. In other organisms,
larval period is positively correlated with size at meta-
morphosis, and it has been hypothesized that larval period
and the stability of larval habitat should be positively re-
lated (Petranka and Sih, 1987; Travis et al, 1987; New-
man, 1988). Applied to Balanus amphitrite, this hypoth-
esis would require that the Salton Sea be a more stable
environment for larvae than coastal bays and harbors.
Effectively this would mean that the negative slope of the
relationship between larval duration and successful set-
tlement would be less extreme for larvae in the Salton
Sea. This seems possible given that predation on larvae,
interspecific competition among larvae, maximum dis-
tance from hard substrate, and the intensity of storms and
currents advecting larvae away from favorable areas for
settlement, should all be less for larvae in the Salton Sea.
One of many alternatives to the preceding hypothesis
is that longer naupliar periods may be an adaptation to
retard a temperature-driven accelerated development rate
that would be detrimental to larvae in the Salton Sea.
Culture temperature has a dramatic positive effect on the
rate of larval development for Balanus amphitrite (Ritt-
schof, pers. comm.; Raimondi, pers. obs.), and temper-
ature in the Salton Sea during the period of maximum
larval abundance averages between 31 and 36°C (Car-
pelan, 1 96 1 b; Linsley and Carpelan, 1 96 1 ). I noticed many
coastal larvae reared at temperatures above 30°C whose
morphology appeared to be intermediate between naupliar
stages (Raimondi, unpubl. data). Such larvae have no fur-
ther development, and their incomplete metamorphoses
may result from temperature-driven differences in the
maximum rate of development of independent physio-
logical processes. Hence, Salton Sea larvae could show
slower development than coastal larvae when reared at
27-28°C because of adaptations to control development
at31-36°C.
Conclusion
The primary aim of this study was to determine whether
evolutionary change or phenotypic plasticity was respon-
sible for the observed phenotypic divergence of a popu-
lation of Balanus amphitrite recently introduced and iso-
lated in the Salton Sea. Clearly, divergence in the exam-
ined adult characters was due to environmentally induced
plasticity. In contrast, there is strong support for the hy-
pothesis that the observed divergence in larval characters
was due to an evolutionary process, probably selection.
These results are not evidence for a general ontogenetic
difference in the way organisms respond to a changing or
novel environment. I suspect that there are some unex-
amined divergent adult traits between populations that
are underlain by genetic differences, and some divergent
larval ones that are not. However, it is clear that the con-
sideration of divergence between populations is incom-
plete if all life history stages of the organism are not stud-
ied. In the present example I would have found no evi-
dence for genetic divergence between the Salton Sea and
coastal populations of Balanus amphitrite if only the adult
morph had been studied. My final comment is a precau-
tionary one. There has been an historic fascination with
examining the causes of phenotypic divergence in isolated
populations by considering them as experimental popu-
lations. Perhaps this is because they resemble experimental
treatments on a larger scale (both temporal and spatial)
and with more ecological realism than is possible in ma-
nipulations. This is flawed thinking, because there is no
PLASTICITY AND EVOLUTION
219
provision for eliminating alternative hypotheses. For Ba-
lanus amphitrite isolated in a novel environment, the Sal-
ton Sea, phenotypic divergence was the expected result
of either of two processes: evolution of phenotypic plas-
ticity. Only through experimental manipulations could
the responsible process be determined, and then only on
a trait-by-trait basis.
Acknowledgments
I am particularly thankful to Dr. Daniel Morse for his
intellectual and financial support in all phases of this re-
search. This study would not have been done without his
invaluable contributions. I also thank C. Amsler, M. Carr,
J. Connell, A. Constable, J. Endler, M. Hart, M. Keough,
C. Lively, D. Macmillan, C. and D. Reed, S. Schuster, J.
Smissen, and D. Stellar for valuable discussions or labo-
ratory assistance. I also thank D. Rittschof and A. Schmidt
of Duke University Marine Laboratory for their expertise
and assistance in cultivation of larvae. This research was
supported in part by a grant to Dr. Daniel Morse by the
Oceanic Biology Program of the U. S. Navy Office of Na-
val Research (Grant # N00014-88-K-0288).
Literature Cited
Anonymous. 1989. In The Los Angeles Times. April 3, 1989.
Atchley, W. R., C. T. Gaskins, and D. Anderson. 1976. Statistical
properties of ratios. I. Empirical results. Syst. Zoo/. 25: 137-148.
Avise, J. C., J. J. Smith, and F. J. Ayala. 1975. Adaptive differentiation
with little genie change between two native California minnows.
Evolution 29:411-426.
Berven, K. A., D. E. Gill, and S. J. Smith-Gill. 1979. Countergradient
selection in the green frog. Rana clamitans. Evolution 33: 609-623.
Bradshaw, A. D. 1965. Evolutionary significance of phenotypic plas-
ticity in plants. Adv. Genet 13: 1 15-155.
Carlton, J. T. 1985. Transoceanic and interoceanic dispersal of coastal
marine organisms: the biology of ballast water. Oceanogr. Mar. Biol
Anna. Rev. 23:313-371.
Carpelan, L. H. 1961a. History of the Salton Sea. State of California
Department of Fish and Game Fish Bulletin 113: 9-13.
Carpelan, L. H. 1961b. Physical and chemical characteristics (of the
Salton Sea). State oj California Department of Fish and Game Fish
Bulletin 113: 17-32.
Chamberlin, G. J., and D. G. Chamberlin. 1980. Colour: its Measure-
ment. Computation and Application. Heyden & Son. London.
Cockerall, T. D. A. 1945. The Colorado desert of California: its origin
and biota. Transactions Kans. Acad. Sci. 48( 1 ): 1 -39.
Connell, J. H. 1980. Diversity and the coevolution of competitors, or
the ghost of competition past. Oikos 35: 131-138.
Day, R. \\ ., and G. P. Quinn. 1989. Comparisons of treatments after
an analysis of variance in ecology. Ecol. Monogr. 59: 433-463.
Dunlop, \V. C., B. E. Chalker, and J. K. Oliver. 1986. Bathymetric
adaptations of reef-building corals at Davies reef. Great Barrier Reef,
Australia. III. UV-B absorbing compounds. J. Exp. Mar. Biol. Ecol.
104: 239-448.
Endler, J. A. 1986. Natural Selection in the Wild. Princeton University
Press, Princeton, New Jersey.
Falconer, D. S. 1989. Introduction to Quantitative Genetics. Third edi-
tion. Longman Scientific & Technical, New York.
Flowerdew, M. W. 1985. Indices of genetic identity and distance in
three taxa within the Balanus amphitrite Darwin complex (Cirripedia,
Thoracica). Crustaceanu 49: 7-15.
Gould, S. J., and R. F. Johnson. 1972. Geographic variation. Anna.
Rev. Ecol. Syst. 3: 457-498.
Gould, S. J., and R. C. Lewontin. 1979. The spandrels of San Marco
and the Panglossian paradigm: a critique of the adaptationist pro-
gramme. Proc. R. Soc. Lond. (B) 205: 581-598.
Henry D. P. and P. A. Mclaughlin. 1975. The barnacles of the Balanus
amphitrite complex (Cirripedia, Thoracica). Zoo/. ( 'erh. 141: 3-254.
Hilton, J. 1945. Where barnacles grow on the sage. Desert Magazine
8(5): 4-6.
Ireland, C., and P. J. Scheuer. 1979. Photosynthetic marine molluscs:
in v/vol4C incorporation into metabolites of the sacoglossan Pla-
cobranchus ocellatus. Science 205: 922-923.
Jokiel, P. J. 1980. Solar ultraviolet radiation and coral reef epifauna.
Science 207: 1069-1071.
Jerlov, N. G. 1950. Ultra-violet radiation in the sea. Nature 166: 1 1 1-
112.
Kornerup, A., and J. H. Wanscher. 1978. Melhuen Handbook of Colour.
3rd edition. Eyre Methuen, London.
Lande, R. 1979. Quantitative genetic analysis of multivariate evolution
applied to brain: body size allometry. Evolution 33: 402-416.
Lande, R. 1980. Genetic variation and phenotypic evolution during
allopatric speciation. Am. Nat. 116: 463-479.
Lande, R. 1982. A quantitative genetic theory of life history evolution.
£«>/ogy63:607-615.
Lewontin, R. C. 1957. The adaptation of populations to varying en-
vironments. Cold Springs Harbor Symp. Quant. Biol. 22: 395-408.
Linsley, R. H., and L. H. Carpelan. 1961. Invertebrate fauna (of the
Salton Sea). State oj California Department of Fish and Game Fish
Bulletin 113:43-61.
Lynch, M. 1984. The limits to life history evolution in Daphnia. Evo-
lution 38: 465-482.
Macdonald. S., C. C. Chinnappa, and D. M. Reid. 1988. Evolution of
phenotypic plasticity in the Slellaria longipes complex: comparisons
among cytotypes and habitats. Evolution 42: 1036-1046.
Nei, M. 1972. Genetic distances between populations. Am. Nat. 106:
283-292.
Newman, R. A., and D. P. Abbott. 1980. Cirripedia: the barnacles. Pp.
504-535 in Intertidal Invertebrates of California. R. H. Morris,
D. P. Abbott, and E. C. Haderlie, ed. Stanford University Press, Stan-
lord, California.
Newman, R. A. 1988. Adaptive plasticity in development ofScaphiopus
couchii in desert ponds. Evolution 42: 774-783.
Patel, B., and D. J. Crisp. 1961. Relation between the breeding and
moulting cycles in cirripedes. Crustaceana 2: 89-107.
Petranka, J. W., and A. Sih. 1987. Habitat duration, length of larval
period, and the evolution of complex life cycle of a salamander, Am-
bysloma texanum. Evolution 41: 1347-1356.
Richardson, B. J., P. R. Baverstock, and M. Adams. 1986. Allozyme
Electrophoresis: A Handbook for Animal Systematic! and Population
Studies. Academic Press Australia, Sydney.
Rittschof, D., E. S. Branscomb, and J. D. Costlow. 1984. Settlement
and behavior in relation to flow and surface in larval barnacles. Bal-
anus amphitrite Darwin. J. Exp. Mar. Biol. Ecol. 82: 131-146.
Rogers, F. L. 1949. Three new species of Balanus amphitrite from
California. / Entomol. Zoo/. 41: 23-32.
Roughgarden, J. R. 1983. Competition and theory in community
ecology. Am. Nat. 122: 583-601.
Schlichting, C. D. 1986. The evolution of phenotypic plasticity in plants.
Annu. Rev. Ecol. Svst. 17: 667-693.
220
P. T. RAIMONDI
Schmalhausen, I. I. 1949. The Theory of Stabilizing Selection. Blak-
iston, Philadelphia.
Smith-Gill, S. J. 1983. Developmental plasticity: developmental con-
version versus phenotypic modulation. Am. Zool. 23: 47-55.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry, Second edition. W. H.
Freeman. San Fransisco, California.
Strathmann, M. F. 1987. Reproduction and Development of Marine
Invertebrates of the Northern Pacific Coast. University of Washington
Press, Seattle, Washington.
Templeton, A. R. 1981 . The evolution of life histories under pleiotrophic
constraints and r-selection. Theor. Popul. Biol. 18: 279-289.
Travis, J., S. B. Emerson, and M. Blouin. 1987. A quantitative-genetic
analysis of larval life history traits in Hyla crucifer. Evolution 41:
145-156.
West-Eberhard, M. J. 1989. Phenotypic plasticity and the origins of
diversity. Anna. Rev. Ecol. Syst. 20: 249-278.
Williams, G. C. 1966. Adaptation and Natural Selection. Princeton
University Press, Princeton, New Jersey.
Yentsch, C. S., and C. M. Yentsch. 1982. The attenuation of light by
marine phytoplankton with specific reference to the absorbtion of
near-UV radiation. Pp. 691-706 in The Role of Solar Ultraviolet
Radiation in Marine Ecosystems. J. Calkins, ed. Plenum Press, New
York.
Underwood, A. J. 1990. Experiments in ecology and management: their
logics, functions, and interpretations. Aitst. J. Ecol. 15: 365-389.
Reference: Biol. Bull 182: 221-230. (April. 1992)
Intercolony Coordination of Zooid Behavior and a New
Class of Pore Plates in a Marine Bryozoan
DANIEL F. SHAPIRO
Section of Ecology and Systematics, Cornell University. Ithaca. New York. 14853.
Abstract. This paper describes a mixed allorecognition
interaction between adjoining colonies of the encrusting
cheilostome bryozoan Membranipora membranacea, in
which characteristics of both intercolony fusion and in-
tercolony rejection occur simultaneously. Intercolony co-
ordination of zooid behavior was assayed by applying
electrical stimuli to one colony of a colony pair while
observing the behavior of the adjoining colony. Retraction
of feeding structures (lophophores) by the unstimulated
colony indicated intercolony coordination of behavior.
Naturally occurring and artificially created pairs of geno-
typically identical and genotypically distinct colonies were
examined. Additionally, colony borders were examined
for the presence of pore plates, structures that physiolog-
ically link zooids within colonies. Contact between ge-
netically identical colonies (isocontact) always resulted in
a characteristic border morphology, characteristic pore
plates, and intercolony coordination of zooid behavior.
Contact between genotypically distinct colonies (allocon-
tact) always resulted in a characteristic border morphology
and in the formation of characteristic pore plates of a type
never before described. However, only colonies that were
young when they first came into contact showed coordi-
nated behavior. Intercolony coordination of zooid be-
havior is probably the result of neural connections made
through pore plates. Intercolony behavioral coordination
between young genotypically distinct colonies is peculiar,
because the colonies simultaneously show characteristics
of physiological integration (coordinated behavior) and
tissue rejection (borders and pore plates characteristic of
contact between genetically distinct tissues). This inter-
action shows that the presence of the morphological char-
acteristics of intercolony rejection does not always imply
a lack of physiological integration between colonies.
Received 4 June 1990; accepted 13 January 1992.
Introduction
Colonial marine invertebrates such as sponges, cnidar-
ians, bryozoans, and ascidians are capable of indetermi-
nate asexual growth. As a result, contact between con-
specific and heterospecific colonies is extremely common
on most marine hard substrata where space is limiting
(Dayton, 1971; Stebbing, 1973a; Jackson, 1977; Osman,
1977). Many of these colonial invertebrates have highly
discriminating immune systems capable of allorecogni-
tion— the ability to distinguish between genetically iden-
tical and genetically distinct tissue (for review see Gros-
berg, 1988). If genotypically identical, or closely related
(e.g.. sibling) colonies come into contact, they commonly
fuse into a single colony. If genotypically distinct colonies
come into contact, tissue rejection typically follows, and
fusion does not occur (Sabbadin, 1982; Scofield et ai,
1982; Chancy, 1983; Rinkevich and Loya, 1983a; Shenk
and Buss, 1991).
Recent work describing allorecognition responses of
colonial marine invertebrates has revealed a diversity of
interactions ranging from intercolony fusion to intercol-
ony rejection. Colonies of the hydroid Hydractinia sym-
biolongicarpm may fuse permanently, fuse and then later
reject, or reject with the subsequent production of ag-
gressive hyperplasitic stolons (Buss and Grosberg, 1990;
Shenk and Buss, 1991). In ascidians, allorecognition re-
sponses include permanent fusion, fusion followed by
separation, fusion followed by complete resorption of one
colony, rejection with little further interaction, and rejec-
tion with necrosis of the tissues of one or both colonies
(Koyama and Watanabe, 1982; Scofield and Nagashima,
1983; Rinkevich and Weissman, 1987, 1989).
All of the above examples involve either different in-
tensities of rejection or a temporal separation between
fusion and rejection. This paper describes a mixed inter-
action between colonies of the encrusting cheilostome
221
222
D. F. SHAPIRO
bryozoan Membranipora membranacea involving simul-
taneous evidence of physiological fusion and tissue rejec-
tion. Zooids within bryozoan colonies are physiologically
integrated through a nerve net that traverses the calcified
zooidal walls through pore plates (Thorpe et al, 1975;
Lutaud, 1977, 1979), distinctive structures in the zooidal
wall where there is a concentration of several open pores
(Silen, 1944; Banta, 1969). The most obvious display of
physiological integration of zooids within a bryozoan col-
ony is the coordination of the lophophore retraction re-
sponse. In response to a localized disturbance to one or
a few zooids, all of the zooids within a colony simulta-
neously retract their feeding structures (lophophores). I
have observed that when genotypically distinct colonies
of M. membranacea come into contact, intercolony co-
ordination of lophophore retraction is frequently ob-
served. Yet, the intercolony borders of these same colonies
show no morphological characteristics effusion.
Because of the mixed nature of this interaction, I will
avoid the terms fusion and rejection. Fusion in bryozoans
is commonly associated with physiological integration
(Stebbing, 1973b; Humphries. 1979; Nielsen. 1981 Cha-
ncy, 1983); consequently, the term fusion could also be
applied to colonies that show physiological integration,
but lack any morphological characteristics of fusion. To
avoid this ambiguity, I will refer to contact between ge-
netically distinct tissues as "allocontact". and 1 will refer
to contact between genetically identical tissues as "iso-
contact." The physiological consequences and morphol-
ogy of these interactions can then be described separately.
M. membranacea occurs naturally in dense monospe-
cific populations where contact between conspecifics is
extremely common, if not unavoidable. Larvae of M.
membranacea disperse in the plankton for up to four
weeks (Yoshioka. 1982), thus naturally settled adjoining
colonies are unlikely to be siblings. Consequently, the
majority of intercolony interactions are between unrelated
colonies. However, contact between genotypically iden-
tical tissues occasionally occurs when a single colony grows
into contact with itself after either growing around some
object or fi ;on resulting from damage to the colony (pers.
obs.). I w 1) examine how intercolony coordination is
related to ; ^ size and age at which genotypically distinct
colonies t/.st come into contact, (2) compare the mor-
phology of the borders between genotypically distinct col-
onies to those between genotypically identical colonies,
and (3) examine both types of borders for pore plates
that could facilitate intercolony coordination of zooid be-
havior.
Materials and Methods
Animal collection
Research was conducted at Friday Harbor Laboratories
(FHL), San Juan Island. Washington, and at the Univer-
sity of California, Los Angeles. At FHL, colonies ofAfem-
branipora membranacea that had settled on black acrylic
panels suspended from the FHL dock, as well as colonies
collected from the field, were used in this study. Colonies
were collected from the field by haphazardly selecting
bryozoan-encrusted blades of the kelp Laminaria sp. from
Turn Island and transporting them back to FHL where
the kelp blades were hung from the FHL dock. In Cali-
fornia, M. membranacea colonies were collected from kelp
beds off the coast of Malibu, California. Bryozoan en-
crusted blades of the kelp Macrocystis pyrifera were hap-
hazardly removed from the upper parts of kelp fronds on,
and just below, the surface of the water. Blades were then
transported back to the laboratory where they were main-
tained in a recirculating seawater system.
Inicrcolony coordination of lophophore retraction
To ensure that a given intercolony border was between
two colonies descended from different larvae rather than
previously separated parts of a colony decended from a
single larva, I used only colony pairs for which I could
locate both ancestrulae. The ancestrula is a pair of mor-
phologically distinct zooids that develop from the larva
after settlement and metamorphosis (Fig. 1 A). Ancestrular
zooids are easily distinguished from younger asexually
produced zooids because they are rounder, more heavily
calcified, and together are distinctively heart shaped (Fig.
1 B). Unless indicated otherwise, whenever I mention col-
ony pairs, I will be referring to pairs of colonies descended
from separate larvae.
To test for intercolony coordination of lophophore re-
traction, I stimulated colony pairs electrically. A stimulus
was applied to one of the two colonies. A colony-wide
lophophore retraction response in the adjoining unstim-
ulated colony was used as an indication of intercolony
behavioral coordination. Electrical stimuli were applied
with an electrode placed on the surface of the colony. All
stimuli were at, or just above, the threshold stimulus (a
single square pulse between 5 and 10 volts for 5-10 ms)
required to elicit a colony-wide lophophore retraction re-
sponse. In addition to electrical stimuli, mechanical stim-
uli were applied to pairs of very small colonies (less than
10 mm2) to eliminate the possibility that intercolony co-
ordination was an artifact resulting from electrical con-
duction of the stimulus through the water or across the
colony surface. Mechanical stimuli were applied by lightly
touching a dissecting needle to one of the colonies on the
edge opposite the intercolony border.
To determine whether the non-stimulated colony of a
pair of behaviorally coordinated colonies was responding
to the physical retraction of the lophophores of the ad-
joining colony, I retested 20 coordinated colony pairs after
first making a fine cut with a razor blade along the border
INTERCOLONY COORDINATION OF ZOOID BEHAVIOR
223
between the adjoining colonies. Cuts were made so that no
lophophores along the intercolony borders were damaged.
To determine whether coordination was bidirectional,
a stimulus was applied to one colony of a pair until I had
obtained 20 behaviorally coordinated and 20 non-coor-
dinated pairs. A second stimulus was then applied to the
other colony of each pair.
The frequency of intercolony behavioral coordination
in a natural population of M. membranacea was measured
at Friday Harbor by sampling three blades ofLaminaria.
Both sides of 5 X 10 cm rectangles were censused 5 cm
from each edge of the blade at 25, 50, 75, 100, and 125
cm from the base of each blade (where the stipe meets
the blade). Each colony was recorded as being solitary or
in contact with other colonies. If a colony was in contact
with another colony, it was tested for intercolony coor-
dination of lophophore retraction. In all, 1301 colonies
were sampled.
Intercolony coordination and size at first contact
To determine the relationship between colony size at
first contact and intercolony coordination, 92 pairs of M.
membranacea colonies were cultured on black acrylic
panels in Friday Harbor. Panels were cleared at least once
a week of all other organisms. Each colony monitored
was in contact with only one other colony. The size of
each colony at the time of first intercolony contact was
determined by tracing each colony on acetate paper and
calculating the area of the tracing using a video-integrated
image analysis system. Following contact, all colony pairs
were tested for intercolony coordination of lophophore
retraction one to three times each week for five weeks.
Additional data on the relationship between intercolony
coordination and colony size at first contact were obtained
for M. membranacea colonies in California. Densities of
M. membranacea in California tend to be higher than in
Friday Harbor (pers. obs.). As a result, data could be ob-
tained for adjoining colonies that were typically smaller
at first contact than those observed in Friday Harbor. In
all, 230 colony pairs were selected from 10 different Mac-
rocystis blades. Colonies were selected to give a maximum
range of values for size at first contact. Because of high
colony density, colony "pairs" were sometimes in serial
contact with other colonies (forming linear groups of 3,
4, or more colonies). However, no colony was ever in
contact with more than two other colonies, and a single
colony was never used more than once. Colonies were
examined using a dissecting microscope, and all mea-
surements were made with an ocular micrometer.
Because I was unable to culture colonies in California,
direct measurements of colony size at first contact were
not possible. Instead, I estimated colony size at first con-
tact by measuring the intercolony ancestrula distance (Fig.
Figure 1 . A. The founding ancestrula of a colony shortly after larval
settlement and metamorphosis. B. Ancestrulae and intercolony border
of a pair of colonies that have grown into contact. Small bubble-like
structures visible along the intercolony border are allocontact pore plates.
Abbreviation: a, ancestrula. Size bars = 0.5 mm.
3). Because the ancestrula marks the site of larval settle-
ment and metamorphosis, I assumed that the distance
between the ancestrulae of two colonies would be directly
correlated to the size of the colonies at first contact. Ad-
ditionally, it seemed likely that colonies would not become
coordinated immediately upon contact, but would instead
require a period of time for the formation of intercolony
physiological connections. Consequently, for each colony
pair I also estimated how long colonies had been in contact
by measuring the intercolony border length (Fig. 3). Be-
cause the length of the border between colonies increases
as both colonies grow, I assumed that the length of the
intercolony border would be directly correlated to how
long the colonies had been in contact. After making these
measurements, colonies were tested for intercolony co-
ordination of lophophore retraction.
Transplant experiment
Although unlikely, I cannot be sure that naturally set-
tled adjoining colonies are not genetically similar siblings
224
D. F. SHAPIRO
that have settled in close proximity. To determine whether
behavioral coordination can occur between colonies that
are clearly not siblings, 1 paired M. membranacea colonies
from Turn Island with colonies from Rocky Point, San
Juan Island, a site approximately 10 miles northwest from
Turn Island. Bryozoan encrusted blades of the red alga
Iridea were collected from the two sites. I removed 48
small colonies (<25 mm2) from the algal blades by gently
stretching the blade until the colony detached. Twenty-
four colony pairs, each consisting of one colony from Turn
Island and one colony from Rocky Point, were then placed
on acrylic panels. After 24 h, colonies had attached to the
panels that were subsequently suspended below the FHL
docks. Following contact, all colony pairs were tested for
intercolony coordination of lophophore retraction twice
each week for four weeks.
Size reduction experiment: allocontact and isocontact
To distinguish the effects of colony age from those of
colony size and to establish unambiguous examples of
isocontact between completely separated parts of a single
colony, I reduced large colonies growing on acrylic panels
at Friday Harbor to pairs of smaller subcolonies. Using a
razor blade to cleanly cut a square of the appropriate size
in the colony. I created pairs of either small or large square
subcolonies that were 16 mm2 or 100 mm2, respectively.
All other parts of the colony were then scraped off the
panel with a small spatula. A 1-mm strip of space was
also scraped between each colony pair.
Allocontact pairs were created by making subcolonies
on both sides of the intercolony border between pairs of
non-coordinated colonies (after testing for behavioral co-
ordination). In all. eight small and seven large allocontact
pairs were established. Isocontact pairs were created by
reducing single colonies into two smaller subcolonies. In
all, seven small and eight large isocontact pairs were es-
tablished. In addition to providing an unambiguous ex-
ample of isocontact, this latter treatment also served as a
control for possible effects of damage on the establishment
of behavioral coordination, because adjoining parts of a
single colony should become physiologically integrated
when they meet. Regeneration and growth of the cut bor-
ders was rapid; all colony pairs had grown back into con-
tact in approximately a week. After subcolonies had grown
into contact, I tested for intercolony behavioral coordi-
nation.
Pore plates
A scanning electron microscope was used to examine
isocontact borders (n = 2) and allocontact borders of co-
ordinated (n = 2) and non-coordinated (n = 2) colony
pairs for the presence of pore plates. Colonies growing on
Laminuria were collected at Fridav Harbor. For isocon-
tact, only single colonies that had grown around some
object and back into contact with itself were used; for
allocontact, only colonies with both ancestrulae present
were used. Colonies were prepared by dissolving away the
tissues of colonies in 2.5% sodium hypochlorite for 12 h
to expose the calcium carbonate skeleton.
Isocontact and allocontact borders of naturally occur-
ring colonies were also examined histologically for pore
plates. I examined isocontact borders (n = 3) and allo-
contact borders between behaviorally coordinated (n = 6)
and non-coordinated (n = 6) colony pairs collected in
California. Approximately 2-3 mm long sections of bor-
ders, along with the kelp substrate, were removed with a
razor blade. Samples were first fixed in 3% glutaraldehyde
in 0.1 M sodium cacodylate buffer. pH 7.4 for 1 h, and
then in 4% osmium in 0. 1 M sodium cacodylate buffer
for an additional hour. Samples were then dehydrated in
a graded series of ethanol dilutions, treated with propylene
oxide, and infiltrated overnight in Medcast low viscosity
embedding medium. After polymerizing overnight at
70°C, samples were sectioned (approximately 3 ^ thick)
and viewed using a light microscope.
Results
Intercolony coordination of lophophore retraction
A cut between behaviorally coordinated colony pairs
always completely eliminated intercolony coordination
of lophophore retraction. Thus, colonies were not re-
sponding to the physical disturbance created by the re-
traction of the lophophores of adjoining colonies.
For all behaviorally coordinated colony pairs tested,
intercolony coordination was always bidirectional. Stim-
ulation of either colony resulted in a colony-wide lo-
phophore retraction response in the non-stimulated col-
ony. Unstimulated colonies of non-coordinated pairs al-
ways failed to respond regardless of which colony was
stimulated. No colony pairs were found in which infor-
mation flow was unidirectional.
Intercolony coordination of behavior is frequently ob-
served in natural populations. Of the 1301 colonies sam-
pled from Laminaria blades, 568 (44%) were in contact
with another colony. Of these, 408 (72%) were behavior-
ally coordinated with at least one neighbor.
Intercolony coordination and size at first contact
Intercolony coordination of zooids was observed most
frequently when two colonies were small at the time of
first contact (Fig. 2). When the areas of each colony in a
pair at the time of first contact were summed, the com-
bined area of colony pairs with coordinated behavior (n
= 15; mean = 1.02 cm2, S.D. = 1.51) was significantly
smaller than the combined area of colony pairs that were
INTERCOLONY COORDINATION OF ZOOID BEHAVIOR
225
«^
19. fl-
^
O
it. 0-
O
w
O
^
6.5-
3
8
3.5-
1-5-
nO*
0 00
^ ^O O O o
O ^D O tt"C O
u.
o
OQD $* Q^S> ° °
111
0.5-
*^^fc S^>° o° ° °
«
t * *O
0.5 1.5 3.5 6.5 11.0 19,0
SIZE OF COLONY 1 (cnf )
Figure 2. Colony size at first contact and intercolony behavior for
92 colony pairs cultured on acrylic panels. Each colony is plotted by the
size of each colony in the pair at the time of initial intercolony contact.
One colony of each pair was arbitrarily designated colony I and the other
colony 2; + = coordinated colony pair, o = non-coordinated colony
pair. Note that data are plotted on a logarithmic scale.
not coordinated (n = 77; mean = 3.05 cm2, S.D. = 3.76;
Mest of In transformed data, P < 0.001). In all pairs that
became behaviorally coordinated, there was a short period
(approximately a week) following initial contact during
which colonies were not behaviorally coordinated. Co-
ordinated behavior of the one outlying pair in Figure 2
was observed on only a single occasion, suggesting either
human error or that the pair was anomalous.
There was a significant relationship between intercolony
coordination and both the estimated size at first contact
(intercolony ancestrula distance; X2 = 46.82, P < 0.0001)
and the estimated length of time in contact (intercolony
border length; x~ = 22.22, P < 0.0001) for colony pairs
from California. Data were analyzed using multiple lo-
gistic regression with intercolony ancestrula distance and
intercolony border length as independent variables and
behavior, coordinated or not coordinated, as the binary
dependent variable. As the estimate of colony size at first
contact increased, the probability of intercolony coordi-
nation decreased (Fig. 3). Although many colony pairs
with small intercolony ancestrula distances (e.g. < 1 mm)
were not coordinated, the majority of these also had small
intercolony border lengths relative to the coordinated col-
ony pairs.
Transplant experiments
Of the 24 colony pairs composed of one colony from
Turn Island and one colony from Rocky Point, 20 re-
mained attached to the acrylic panels and grew into con-
tact. Of these, 13 (65%) showed intercolony coordination
of zooid behavior within two weeks, thus demonstrating
that coordinated behavior can occur between colonies that
are clearly not related. For all colony pairs that became
coordinated, there was a brief period (approximately one
week) following initial intercolony contact when colonies
were not coordinated.
Size reduction experiment: allocontact and isocontact
In the size reduction experiment, none of the allocon-
tact colony pairs became behaviorally coordinated, re-
gardless of size. Thus, if colonies are genotypically distinct,
age rather than size appears to be the most important
factor determining whether intercolony coordination oc-
curs. All isocontact colony pairs did become behaviorally
coordinated, regardless of size.
Isocontact borders were morphologically distinct from
allocontact borders. Isocontact borders were straight and
fully calcified, and zooids distal to the area of first contact
aligned to form a single growing edge (Fig. 4A). Allocon-
tact borders were clearly distinct from isocontact borders.
Allocontact borders were not as straight as isocontact bor-
ders and were uncalcified or only lightly calcified, and
each colony maintained a separate growing edge (Fig. 4B).
Pore plates
All isocontact and all allocontact borders (both from
coordinated and non-coordinated colony pairs) contained
structures (Figs. 5, 6) that clearly resemble the previously
described bryozoan pore plates, transverse pore plates,
lateral pore plates, and fusion pore plates (Silen, 1944;
Banta, 1969; Chancy, 1983). Transverse and lateral pore
plates are located respectively in the transverse and lateral
zooidal walls that separate adjoining zooids within the
same colony (Figs. 5A, B; 6A). Fusion pore plates are
found in the walls between two colonies that have fused
into a single colony (Chancy, 1 983). Pore plates are round
(lateral and fusion pore plates) to elliptical (transverse pore
Intercolony Ancestrula Distance
33.0
S 190
|f 11.0
m E
6.5
3.5 -
35
8§
£ '5
I 0.5
Intercolony Border Length
O
— I —
1.5
— l —
3.5
— 1 —
6.5
— I
11.0
— I
19.0
— I
33.0
INTERCOLONY ANCESTRULA DISTANCE (mm)
Figure 3. Intercolonial ancestrula distance, intercolonial border
length and intercolonial behavior for 230 colony pairs from Macrocyslis
blades. For each colony pair intercolony ancestrula distance (estimate
of size at first contact) is plotted against intercolony border length (es-
timate of time since first contact); + = coordinated colony pair, o = non-
coordinated colony pair. Note that data are plotted on a logarithmic
scale.
226
D. F. SHAPIRO
-
•vs^JSSs^Vv^^&MV*?
Figure 4. A. Isocontact border between genotypically identical tissues
of a colony that has grown around another colony and back into contact
with itself. B. Allocontact border between genotypically distinct colonies.
Abbreviations: ab, allocontact border; ib, isocontact border; Size bars
= 1.0 mm.
plates) in shape and slightly raised to form a perforated
calcium carbonate dome or lens, the base of which is at-
tached to the zooidal wall (Fig. 5A-C).
Pore plates found in isocontact borders, herein referred
to as "isocontact pore plates" (Figs. 5C, 6B), were similar
to lateral pore plates in that they consisted of a single
round perforated dome. However, whereas lateral pore
plates tended be of a uniform size and regularly spaced
in lateral walls, isocontact pore plates were variable in
size and occurred irregularly, occasionally in groups, in
the walls formed between genotypically identical colonies.
Pore plates found in allocontact borders, herein referred
to as "allocontact pore plates" (Figs. 5D; 6C, D) were
found in the borders between both coordinated and non-
coordinated colonies. Whereas all previously described
pore plates consist of a single perforated calcium carbonate
dome, allocontact pore plates were composed of two per-
forated calcium carbonate domes placed base to base
forming a single sphere embedded in the intercolony bor-
der. Allocontact pore plates also differed from other pore
plates in that they generally had three or fewer pores. In
contrast, other types of pore plates generally had four or
more pores. There were no obvious morphological dif-
ferences between allocontact pore plates of coordinated
and non-coordinated colonies.
Discussion
The results of this study show that allorecognition re-
sponses following contact between colonies of the bry-
ozoan Membranipora membranacea vary depending on
the genetic similarity and age of interacting colonies.
Contact between genetically identical colonies is always
characterized by an isocontact border, isocontact pore
plates, and coordinated behavior of zooids. Contact be-
tween genotypically distinct colonies is always character-
ized by allocontact borders and allocontact pore plates.
However, only colonies that are young when they first
come into contact, show coordinated behavior.
Intercolony coordinated behavior appears to be the re-
sult of intercolony neural integration. Thorpe et al. (1975)
demonstrated the presence of electrical signals that con-
ducted across colonies of M. membranacea at the same
rate as the spread of lophophore retractions. Electrical
signals similar to those described by Thorpe et al. (1975)
have been found to pass between behaviorally coordinated
colonies but not between non-coordinated colonies
(Shapiro and Mackie, unpub. data), providing direct ev-
idence of intercolony neural linkage.
The presence of pore plates provides morphological ev-
idence for intercolony neural linkage. The time required
for the formation of isocontact or allocontact pore plates
following initial intercolony contact would explain why
colonies did not become coordinated immediately upon
contact and why colonies with short intercolony border
lengths did not show coordinated behavior. However, the
presence of allocontact pore plates does not necessarily
indicate behavioral coordination because allocontact pore
plates were also found between non-coordinated colony
pairs. Thus, there may be morphological differences on a
finer scale (e.g., presence or absence of functional nerves)
between the allocontact pore plates of behaviorally co-
ordinated and non-coordinated colonies.
Allocontact pore plates represent a new, morphologi-
cally distinct class of pore plates never before described
in the Bryozoa. This is the first time pore plates between
unrelated bryozoan colonies have been described. Chancy
1NTERCOLONY COORDINATION OF ZOOID BEHAVIOR
227
Figure 5. Scanning electron micrographs of the different types of pore plates found in Membranipora
membranacea. A. Basal view of calcined zooidal walls showing transverse and lateral pore plates between
zooids within a colony. (lOOx). B. Lateral pore plate between zooids within a colony (500X). C. Isocontact
pore plates (500X). D. Allocontact border showing allocontact pore plates ( 100 • ). Abbreviations: ab. allo-
contact border; ap. allocontact pore plate: Kv. lateral wall: tw. transverse wall.
(1983) examined the borders between unrelated colonies
of the cheilostome bryozoan Thalamoporella califomica,
but found no evidence of pore plates. However, Chancy
( 1983) did find pore plates between sibling colonies of T.
califomica. These pore plates, which he called fusion pore
plates, consisted of a single rather than a double calcium
228
D. F. SHAPIRO
B
Figure 6. Light micrographs of the different types of pore plates found in Membranipora membranacea.
Sections A through C were made parallel to the plane of the colony. A. Transverse and lateral pore plates
between zooids within a colony. B. Isocontact plates. C. Allocontact plate. D. Section perpendicular to
allocontact border and plane of colony showing an allocontact plate. Abbreviations: ab. allocontact border;
ap, allocontact plate; ib, isocontact border; ip, isocontact plate; k, kelp; Iw, lateral zooidal wall; tp, transverse
pore plate; tw, transverse zooidal wall. Size bars = 20 ^m.
carbonate dome and thus resemble the isocontact pore
plates described in this study and not allocontact pore
plates. Additionally, fusion pore plates, like isocontact
pore plates, were variable in size and occurred irregularly
in the walls formed by contact between two colonies. T.
californica larvae settle within hours of release from the
parental colony (Chancy, 1983), thus indicating the po-
tential for substantial inbreeding in natural populations
(Jackson, 1986). Consequently, although sexually pro-
duced, sibling colonies may be nearly genetically identical.
Thus, fusion pores plates are probably the same as iso-
contact pore plates, both being characteristic of contact
between genetically similar tissues.
It is usually assumed that colony pairs that have the
morphological characteristics of fusion are physiologically
integrated, and unfused colonies are not (Humphries,
1979;Stebbing, 1973b; Buss, 1982;Chaney, 1983). How-
ever, assays for physiological integration are rarely per-
formed (Hidaka, 1985; Rinkevich and Loya, 1983a, b).
When Rinkevich and Loya ( 1983a) used SEM to examine
the allocontact borders between colonies of the Red Sea
coral Stylophora pistillata with the morphological char-
acteristics effusion, they found that the colonies were not
physiologically connected. In contrast, this study has
demonstrated that M. membranacea colonies with the
morphological characteristics of rejection can be physi-
INTERCOLONY COORDINATION OF ZOOID BEHAVIOR
229
ologically connected. Thus, unless adequate tests are per-
formed, it may not always be safe to use morphological
evidence of fusion or rejection to imply the presence or
absence of physiological integration.
Allorecognition responses are important in intra- and
interspecific interactions. Rejection responses to contact
with colonies often result in the induction of aggressive
structures used to fight, damage, or surround neighboring
colonies (e.g., Ivker, 1972; Francis, 1973; Rinkevich and
Loya, 1983a; Sebens and Miles, 1988; Harvell and Padilla,
1990). On the other hand, fusion responses may benefit
interacting colonies by increasing competitive ability, in-
creasing fecundity, decreasing probability of mortality, or
decreasing age of first reproduction (Buss, 1982). However,
it may be erroneous to always associate fusion with co-
operation and rejection with aggression. Rinkevich and
Weissman (1987, 1989) found that fusion between geno-
typically distinct ascidian colonies frequently resulted in
partial or total resorption of one of the colonies at a cost
to both colonies. Thus, in this case fusion is apparently
an aggressive interaction. In contrast, the results of the
present study indicate the potential for cooperation be-
tween colonies that do not appear to have fused.
Intercolony behavioral coordination may be an adap-
tation that benefits small colonies by reducing the prob-
ability of mortality. Mortality of many marine inverte-
brates, including bryozoans, is size dependent, with small
colonies having a higher probability of mortality (Jackson,
1985; Yund and Parker, 1989; Harvell el al, 1990). Co-
ordinated behavior between small M. membranacea col-
onies may benefit each colony by enabling colonies to
receive and transmit signals that act as "warnings" of pos-
sible sources of damage or mortality. Such cooperative
behavior is consistent with theory predicting that coop-
eration will be more likely to evolve between sessile or-
ganisms that interact repeatedly (Axelrod and Hamilton,
1981; Buss, 1981).
It could also be argued that intercolony behavioral co-
ordination is a non-adaptive trait that results from the
inability of young colonies to distinguish between genet-
ically identical and genotypically distinct tissues. There
are several examples of colonial marine invertebrates that
will fuse when young but not when older (e.g., Hidaka,
1985:Shenkand Buss. 1991). It is not known what causes
changes in fusibility, although immunological incompe-
tence of young colonies has been suggested (Hidaka,
1985). However, if immature, genotypically distinct col-
onies of M. membranacea were simply treating adjoining
colonies as genotypically identical, intercolony borders
and pore plates should resemble isocontact borders and
isocontact pore plates. Instead, typical allocontact borders
and allocontact pore plates were formed between all ge-
notypically distinct behaviorally coordinated colonies
implying that colonies had recognized their neighbors as
being genotypically distinct.
Acknowledgments
I thank Jim Morin for generously providing lab space
and materials in California and Andrea Huvard for in-
structing me in the techniques of light microscopy. This
manuscript benefited from comments by Liz Francis, Jim
Morin, Drew Harvell, Josh Nowlis, Jordan West, Staci
Eisner, and an anonymous reviewer. Discussions with Liz
Francis were extremely helpful in organizing the final
draft. This research was supported in part by the Lerner-
Gray Fund for Marine Research and NSF-OCE-88 17498
to C. Drew Harvell. I also thank Dennis Willows for pro-
viding space and facilities at Friday Harbor Laboratories.
Literature Cited
Axelrod R., and VV. D. Hamilton. 1981. The evolution of cooperation.
Science 211: 1390-1396.
Banta, VV. C. 1969. The body wall of cheilostome Bryozoa, II. Inter-
zooidal communication organs. / Morphol. 129: 149-170.
Buss, L. W. 1981. Group living, competition, and the evolution of
cooperation in a sessile invertebrate. Science 213: 1012-1014.
Buss, L. W. 1982. Somatic cell parasitism and the evolution of somatic
tissue compatibility. Proc. Natl. Acad. Sci. USA 79: 5337-5341.
Buss, L. W., and R. K. Grosberg. 1990. Morphogenetic basis for phe-
notypic differences in hydroid competitive behavior. Nature 343: 63-
66.
Chancy, H. W. 1983. Histocompatibility in the cheilostome bryozoan
Thalamoporella ca/ifornica. Trans. Am. Microsc. Soc. 102: 319-332.
Dayton, P. K. 1971. Competition, disturbance, and community or-
ganization: the provision and subsequent utilization of space in a
rocky intertidal community. Ecol. Manogr. 41: 351-389.
Francis, L. 1973. Clone specific segregation in the sea anemone An-
thopleura elegantissima. Biol Bull. 144: 64-72.
Grosberg, R. K. 1988. The evolution of allorecognition specificity in
clonal invertebrates. Q. Rev. Biol. 63(4): 377-412.
Harvell, C. D., and D. K. Padilla. 1990. Inducible morphology, het-
erchrony, and size hierarchies in a marine bryozoan. Proc. Natl. Acad.
Sci. USA 87: 508-512.
Harvell, C. D., H. Caswell, and P. Simpson. 1990. Density effects in
a colonial monoculture: experimental studies with a marine bryozoan
(Membranipora membranacea L.). Oecologia 82: 227-237.
Hidaka, M. 1985. Tissue compatibility between colonies and between
newly settled larvae of Pocillopora damicornis. Coral Reefs 4: 111-
116.
Humphries, E. M. 1979. Selected features of growth in Parasmittina
nitida. Pp. 195-2 1 8 in Advances in Bryozoology, G. P. Larwood and
M. B. Abbott, eds. Academic Press, New York.
Ivker, F. 1972. A hierarchy of histo-incompatability in Hydractinia
echinata. Biol. Bull. 143: 162-174.
Jackson, J. B. C. 1977. Competition on marine substrata: the adaptive
significance of solitary and colonial strategies. Am. Nat. Ill: 743-
767.
Jackson, J. B. C. 1979. Overgrowth competition between encrusting
cheilostome ectoprocts in a Jamaican cryptic reef environment. /.
Anim. Ecol. 48: 805-823.
Jackson, J. B. C. 1985. Distribution and ecology of clonal and aclonal
benthic invertebrates. In Population Biology and Ecology of Clonal
230
D. F. SHAPIRO
Organisms. J. B. C. Jackson. L. W. Buss, and R. E. Cook, eds. Yale
University Press. New Haven and London.
Jackson, J. B. C. 1986. Modes of dispersal of clonal and aclonal benthic
invertebrates: consequences for species' distributions and genetic
structure of local populations. Bull. Mar. Sci. 39: 588-606.
Koyama, H., and H. Watanabe. 1982. Colony specificity in the ascidian,
Perophora sagamiensis. Biol Bull 162: 171-181.
Lutaud, G. 1977. The bryozoan nervous system. In The Biology of
Brvo:oans, R. M. Woollacott and R. L. Zimmer, eds. Academic
Press, New York and London.
Lutaud, G. 1979. Etude ultrastructurale du plexus colonial et recherche
de connexions nerveuseo interzoidiales chez le bryozoaire chilostome
Electro pilosa (Linne). Cah. Biol. Mar. 20: 315-324.
Nielson, C. 1981. On morphology and reproduction of Hippodiplosia
insculpta and Fenestnilina malusu (Bryozoa, Cheilostomata). Ophelia
20(1): 91-125.
Osman, R. W. 1977. The establishment and development of a marine
epifaunal community. Ecol. Monogr. 47: 37-63.
Rinkevich, B., and Y. Loya. 1983a. Intraspecific competitive networks
in the Red Sea coral Stylophora pistil/ata. Coral Reefs 1: 161-
172.
Rinkevich, B., and Y. Ixjya. 1983b. Oriented translocation of energy
in grafted reed corals. Coral Reels 1: 243-247.
Rinkevich, B., and I. L. Weissman. 1987. A long-term study on fused
subclones in the ascidian Botryllus schlosseri: the resorption phe-
nomenon (Protochordata: Tunicata). J. Zoo/. Land. 213: 717-
733.
Rinkevich, B., and I. L. Weissmann. 1989. Variation in the outcome
following chimera formation in the colonial tunicate Botryllus
schlosseri. Bull. Mar. Sci. 45(2): 213-227.
Sabbadin, A. 1982. Formal genetics of ascidians. Am. Zool. 22: 765-
773.
Scofield, V. L., J. M. Schlumpberger, L. A. West, and I. L. Weissman.
1982. Protochordate allorecognition is controlled by a MHC-like
gene system. Nature 295: 499-502.
Scofield, V. L., and L. S. Nagashima. 1983. Morphology and genetics
of rejection reactions between oozoids from the tunicate Botryllus
schlosseri. Biol. Bull. 165: 733-744.
Sebens, K. P., and J. Miles. 1988. Sweeper tentacles in a gorgonian
octocoral: Morphological modifications for interference competition.
Biol. Bull. 175: 378-387.
Shenk, M. A., and L. W. Buss. 1991. Ontogenetic changes in fusibility
in the colonial hydroid Hydractinia symbiolongicarpus. J. Exp. Zool.
257: 80-86.
Silen, L. 1944. On the formation of the interzoidal communications
of the Bryozoa. Zool. Bidr. Upps. 22: 433-488.
Stebbing, A. R. D. 1973a. Competition for space between the epiphytes
of Fuc us serratus L. J. Mar. Biol. Assoc. t/A'53: 247-261.
Stebbing, A. R. D. 1973b. Observations on colony overgrowth and
spatial competition. Pp. 173-183 in Living and Fossil Bryozoa.
G. P. Larwood, ed. Academic Press, London and New York.
Thorpe, J. P., G. A. B. Shelton, and M. S. Laverack. 1975.
Electrophysiology and coordinated responses in the colonial bryozoan
Membranipora membranacea (L.). J. Exp. Biol. 62: 115-121.
Yoshioka, P. M. 1982. Role of planktonic and benthic factors in the
population dynamics of the bryozoan Membranipora membranacea.
Ecology 63(2): 457-468.
Yund, P. O., and H. M. Parker. 1989. Population structure of the
colonial hydroid Hydractinia sp. nov. C in the Gulf of Maine. J.
Exp. Mar. Biol. Ecol. 125: 63-82.
Reference: Biol. Bui! 182: 231-240. (April, 1992)
Antipredator Defenses in Tropical Pacific Soft Corals
(Coelenterata: Alcyonacea). I. Sclerites as Defenses
Against Generalist Carnivorous Fishes
KATHRYN L. VAN ALSTYNE1, CHAD R. WYLIE,
VALERIE J. PAUL, AND KAREN MEYER
University of Guam Marine Laboratory, UOG Station, Mangilao, Guam 96923
Abstract. Calcified sclerites are common in many in-
vertebrate species and are frequently used as taxonomic
indicators; however, little is known about the function of
sclerites. To determine whether sclerites could function
as antipredator defenses, we conducted field assays in
which sclerites from the Indo-Pacific soft corals Simdaria
maxima, S. polydactyla, and 5. sp. were incorporated into
an artificial diet and offered to a natural assemblage of
fishes in the field. Sclerites from both the tips and bases
of all three species ofSinularia reduced feeding by a nat-
ural assemblage of generalist carnivorous fishes offGuam
by 27-44%; however, sclerites from the bases of the col-
onies were 18-51% more deterrent than tip sclerites. The
greater effectiveness of sclerites from the bases of the col-
onies was largely attributable to their high concentrations.
Sclerites in the tips of the colonies occurred in mean con-
centrations from 24 to 58% by dry weight and were gen-
erally less than 0.5 mm in length. Sclerites in the bases of
the colonies were larger and occurred in average concen-
trations of 82-88%. Simdaria sclerites were increasingly
effective as feeding deterrents with increasing concentra-
tion at concentrations less than 30-50% by dry weight.
The effectiveness of sclerites as deterrents leveled off at
higher concentrations. Sclerite morphology was also im-
portant in determining antipredator activity. Although
sclerites can play a role in predator deterrence, they can
also function in the structural support of colonies. Thus,
the sizes, shapes, and abundances of sclerites at different
locations within colonies will be determined by their
Received 1 1 July 1991; accepted 23 January 1992.
' Present Address: Department of Biology, Kenyon College, Gambier.
Ohio 43022.
functions at particular locations as well as constraints
upon their use or production.
Introduction
Soft corals are frequently a conspicuous component of
shallow, Indo-Pacific tropical reef communities despite
the abundance of carnivorous fishes. For example, on
Guam, soft corals have been reported to provide ~95%
of the total living animal cover on some reefs (Wylie and
Paul, 1989). On New Guinean reefs, soft corals constitute
approximately 50% of the living cover between depths of
0 and 5 m (Tursch and Tursch, 1982).
The persistence of soft corals and gorgonians in areas
with high levels of predation has previously been attributed
to their production of predator-deterrent secondary me-
tabolites (Coll et ai, 1983; LaBarre et al., 1986; Pawlik
et ai, 1987; Wylie and Paul, 1989); but, soft corals and
gorgonians also produce sclerites or spicules that could
potentially serve as antipredator defenses (Harvell and
Suchanek, 1987; Sammarco et al., 1987; Harvell et al.,
1988; Harvell and Fenical, 1989). Mineral-hardened spic-
ules are common within a number of invertebrate groups
including sponges, cnidarians, platyhelminth worms,
mollusks, echinoderms, and ascidians (Kingsley, 1984).
However, despite the widespread occurrence of sclerites
within marine invertebrates, little is known about the
function of these structures, in particular, their ability to
deter feeding by potential predators.
Only in recent studies has the role of sclerites as anti-
predator defenses been explored. When incorporated into
an artificial diet, sclerites of the gorgonian sea whip Pseu-
dopterogorgoria acerosa deterred feeding by carnivorous
fishes in field assays in Belize (Harvell et al.. 1988). Scler-
231
232
K. L. VAN ALSTYNE ET AL.
ites of the Caribbean gorgonian Gorgonia ventalina deter
feeding by natural assemblages of fishes in the field and
by the gorgonian specialist Cyphoma gibbosum (Van Al-
styne and Paul, in press). Toxicity of soft corals to the
mosquito fish Gambmia affinis was negatively correlated
with the degree of armament of the polyps in the Neph-
thedidae and negatively correlated with the degree of
mineralization of the coenenchyme in 14 species of Sin-
itlaria (Sammarco et a!., 1987). Alcyonarian sclerites are
extremely variable in size and morphology and are fre-
quently used as taxonomic characters (e.g., Bayer, 1956,
1961; Bayer el ai. 1983). Seasonal fluctuations in collagen
levels in the organic matrix of sclerites indicate that these
structures are dynamic and may undergo seasonal cycles
of demineralization and remineralization (Kingsley et ai.
1990).
In this paper, we provide direct experimental evidence
that sclerites from the alcyonian soft corals Sinularia spp.
serve as antipredator defenses to generalist carnivorous
fishes in the field. Soft corals of the genus Sinularia are
widely distributed throughout the Indo-Pacific region
(Verseveldt, 1977, 1980), and are an important compo-
nent of shallow coral reef communities on Guam (Gawel,
1977). Sinularia spp. generally have a lobate morphology
with an upper portion that contains polyps, and a lower
sterile stalk (Verseveldt, 1977). The surface of the lobes
contains small club-shaped sclerites, rods, and spindles
(Verseveldt, 1980). We also document differences in
sclerite morphologies and densities within soft coral col-
onies and explore the consequences of these differences
in deterring feeding by fishes.
^Cocos
Lagoon
Figure 1. Locations of the Fingers Reef and Cocos Lagoon study
sites on Guam.
Tips
Base
Figure 2. Sampling locations within an individual Sinularia colony.
Three transects of six samples each were made from the tips of the colony
to the base of the colony. Transects ran from the most distal tips of the
colonies (location 1 ) to the point of attachment at the base of the colony
(location 6).
Materials and Methods
Studv sites
Sinularia maxima, S. polydactyla. and S. sp. were col-
lected from a patch reef in Cocos Lagoon, Guam, USA
(Fig. 1). This reef has been previously described by Paul
and Van Alstyne (1987) and Wylie and Paul (1989). It is
a small patch reef, ~35 m X 50 m, that is composed
mainly of dead Acropom. Soft corals comprise ~ 10% of
the cover of the reef (Wylie, 1988).
All feeding experiments took place on Fingers Reef in
Apra Harbor, Guam. The reef at Fingers (Fig. 1) is com-
posed of a diverse assemblage of scleractinian corals, as
well as many unidentified species of Sinularia. Numerous
species of herbivorous and carnivorous fish inhabit this
reef. The fishes that were observed to feed on ropes during
feeding experiments included sergeant majors (Abudefduf
spp.), damselfish (Amblyglyphidodon curacao), wrasses
(Cheilinus fasciatm, Gomphosus varius, Halichoeres tri-
macitlalus), and triggerfish (Balistapus undulatiis).
Quantification of sclerite concentrations
Five soft corals each of Sinularia maxima, S. polydac-
tyla, and 5. sp. were collected from Cocos Lagoon. Six
samples of ~2 cm (height) X 1 cm (length) x 1 cm (width)
were removed from each of three "transects" of each col-
ony. The transects ran from the most distal part of the
tips to the point of attachment at the base (Fig. 2). These
samples areas will be referred to as locations #1 (most
STRUCTURAL DEFENSES IN SOFT CORALS
233
distal tip sample) through #6 (base sample closest to the
attachment point of the colony). Eighteen samples were
taken from each of the 15 colonies. Sclerites concentra-
tions from each sample were quantified as described in
Van Alstyne and Paul (in press).
An analysis of variance was used to test for differences
in mean sclerite concentrations among species, among
individuals within a species, among transects within an
individual, and among locations. A mixed model ANOVA
was conducted with the factors being ( 1 ) species, (2) in-
dividuals nested within species, (3) transects nested within
individuals, and (4) locations along the transects from the
tips to the bases of the colonies. The analysis was con-
ducted with an SPSSPC + statistical package.
Measurement ofsderites
Sclerites from the samples taken above were used to
obtain measurements of sclerite sizes within colonies.
Measurements of maximum sclerite length were made
from 100 to 200 sclerites from each sample using a video
image analysis system. All of the sclerites in at least two
randomly selected fields from each of three subsamples
from each sample were measured to the nearest jum at
magnifications ranging from 60X to 500X using a JAVA
computer program.
Sclerite masses were estimated by using a log-log
regression of sclerite masses to lengths. To obtain the
regression, 15-20 sclerites from each species were mea-
sured to the nearest ^m with a video image analysis system
and then weighed to the nearest 0.01 mg. The regression
equations for each species were: S. maxima: mass(g)
= e-26 85 + 3.10-ln[length(^m)] (n = 2Q, r = 0.696, P < 0.05), S.
polvdactvla: mass(g) = e-|880+2 15*'nllensth("m)] (n = 20, r2
= 0.802, P< 0.05), 5. sp: mass(g) = e-2000+241"lnl|i:nPh(''m))
(n = 15, r2 = 0.960, P < 0.05).
Extraction of sclerites for feeding assays
Approximately 1 5 individuals each of 3 soft coral spe-
cies were collected from Cocos Lagoon. Tissues at the
bases of the soft corals were separated from the tips and
lobes (sensu Wylie and Paul, 1989). The polyp-bearing,
top 10-15 cm of the colonies will hereafter be referred to
as the tips (Fig. 2). The pieces from all individuals of a
species were combined and extracted twice in 1 : 1 meth-
ylene chloride:methanol. The organic extracts were re-
moved, and the remaining soft coral pieces were dried,
then digested in bleach until no organic material re-
mained. The sclerites were rinsed five to eight times in
tap water, then once with acetone, and dried. Examination
of sclerites using light microscopy and scanning electron
microscopy showed no signs of sclerite deterioration due
to the extraction process.
Two-choice feeding experiments
Sclerites from soft coral tips and bases were incorpo-
rated into an artificial diet consisting of 2.5 g carrageenan
(Sigma #C-1013), 4 g of paraffin wax, and 70 ml water,
heated in a microwave oven for 75 s. After the wax/car-
rageenan mixture was heated, ~50 g of squid homogenate
(250 ml water:500 g squid) and sclerites were added. Con-
centrations of sclerites used in these feeding experiments
are listed in Table I. These values are within the ranges
of concentrations measured in bases and tips ofSinularia
spp. in this study. Controls consisted of the artificial diet
without sclerites. The diet was then poured into stainless
steel potato sheers that were composed of a 7 X 7 grid of
1 cm X 1 cm squares that were ~ 1 cm high. Prior to the
addition of the diet, black plastic o-rings (3/8 in OD, '/» in
ID) were placed in each cube in the potato sheer. After
the diet had gelled, it was removed from the mold and
attached to ropes to be placed out in the field. The diet
was presented to fishes on 50 cm long pieces of 3 strand,
'A in yellow polypropylene rope in which four 3-cm long
safety pins were attached equidistantly along the top 30-
40 cm of the rope (Fig. 3).
The ropes were placed on the reef in pairs by attaching
them to pieces of coral. All experiments were done on
Fingers Reef in Apra Harbor, Guam, at a depth of 5-8 m.
Pairs of ropes were placed on the reef and removed
when at least four of the eight pieces of artificial diet on
the two ropes had been completely consumed. Each ex-
periment consisted of 17-21 pairs of ropes. Differences
in consumption were compared with a Wilcoxon Signed
Ranks test (a = 0.05).
Feeding experiments at Fingers Reef consisted of the
following paired experiments: ( 1) comparisons of feeding
rates on control diet versus diet containing either tip
sclerites or base sclerites at natural concentrations (Table
I), (2) comparisons of feeding rates on control diet and
diet containing either tip or base sclerites at 10%, 25%, or
50% by dry weight, and (3) comparisons of diet containing
tip sclerites at natural concentrations with diet containing
base sclerites at natural concentrations. Each of the ex-
periments described above was conducted with all three
species of Sinularia.
Table I
Concentrations (as percent of dry weight) of sclerites used
in feeding assays
Tips
Bases
Sinn/aria maxima
31%
81%
S. polydactyla
41%
76%
S. sp.
47%
82%
234
K. L. VAN ALSTYNE ET AL.
Figure 3. Rope used in a feeding assay. The assay consisted of at-
taching an o-ring (C) embedded within a piece of artificial diet (E) to a
safety pin (B). The safety pins were attached to a 0.5 m piece of 3-strand
polypropylene rope (F) and were buoyed with small pieces of neoprene
(D). The ropes were attached in the field to pieces of coral (G). Ten-cm
pieces of labelled surveyor's tape (A) were used to distinguish ropes con-
taining different types of diet. A single rope held only one type of diet.
Four-choice feeding experiments
To determine the relative effects of sclerite morphol-
ogies and concentrations on feeding deterrence, fishes were
offered artificial diets containing two types of sclerites (tip
and base) at two different concentrations (tip concentra-
tions and base concentrations). Thus, fishes were offered
tip sclerites at natural tip concentrations, tip sclerites at
natural base concentrations, base sclerites at natural tip
concentrations, and base sclerites at natural base concen-
trations. Separate experiments were conducted for each
of the three Sinularia species.
Artificial diets were attached to polypropylene ropes
and offered to fishes in the field in the same manner as
described for the two-choice experiments, with the excep-
tion that a set of four ropes were offered instead of a pair
of ropes. Each of the four ropes contained one of the four
diets described above; each of the four types of diet was
presented in each replicate (n = 19 or 20). Ropes were
removed and the number of pieces of each diet remaining
on each rope were counted after fishes had consumed at
least half of the sixteen pieces of diet in the set. Statistical
analyses were conducted with a two-way Kruskal-Wallis
test. Factors used in the analysis were sclerite type and
sclerite concentration.
Results
Distribution and abundance of sclerites
Dry weight concentrations of sclerites differed among
the three species of Sinularia and among locations within
the colonies (Table II). In general, sclerite concentrations
(as % of dry weight) increased from the tips of the colonies
to their bases (Fig. 4). Mean concentrations at the tips of
the colonies ranged from 24% in S. polydactyla to 58%
in 5. sp. Mean concentrations of sclerites in the most
basal portion of the colonies ranged from 82% in S. max-
ima to 88% in S. sp.
The size distributions of sclerites also showed consid-
erable intra-colony variation. Almost all of the sclerites
in the most distal tips of the colonies of the three species
of soft corals were less than 0.5 mm in length (Figs. 5-7).
However, in S. sp., the majority of the mass of the sclerites
in the most distal tips were comprised of sclerites that
were greater than 0.5 mm long; the largest fraction by
mass was 1.0-1.5 mm long (Fig. 5). In the most distal tips
of S. polydactyla, approximately 55% of the sclerite mass
was comprised of sclerites less than 0.5 mm in length (Fig.
6); and, in S. maxima, all of the mass of the sclerites in
the most distal tips was in the <0.5 mm category (Fig. 7).
Larger sclerites increased in abundance in the more
basal portions of the colonies of all three soft coral species
(Figs. 5-7). Even in the most basal samples, sclerites less
than 0.5 mm long were numerically abundant; however,
they made up only a small fraction of the total sclerite
Table II
Ana/vsis of variance of sclerite concentrations in Sinularia spp.
Source
df
SS
MS
Species
2
12,130
6,065
10.39
0.000
Individual (species)
12
8,363
697
1.19
0.288
Transect (individual)
10
6,483
648
1.11
0.355
Location
5
55,625
11,124
19.06
0.000
Error
240
140,071
584
Total
269
222,672
A mixed model ANOVA was conducted with the factors being ( I )
species, (2) individuals nested within species. (3) transects nested within
individuals, and (4) locations along the transects from the tips to the
bases of the colonies.
STRUCTURAL DEFENSES IN SOFT CORALS
235
Sclerite Concentrations
100
80
o
O
60J
40
20-
Tips
Base
• S max sclentes •
S poly sdemes -*- S sp sclenies
Figure 4. Sclerite concentrations (as % of dry weight) within colonies
of Sinn/aria spp. Samples were obtained from five colonies of each of
the three species. The samples correspond with locations 1 (tips) through
6 (base) shown in Figure 2. Horizontal bars represent ± 1 SD.
mass. In the most basal samples, the largest fractions of
sclentes by mass were in the 1.0-1.5 mm size class in S.
maxima (Fig. 7) and in the 1.5-2.0 mm size class in S.
polydactyla (Fig. 6) and S. sp. (Fig. 5). Thus, in all three
species of Sinularia examined in this study, sclentes in-
creased in both size and concentration from the tips of
the colonies to the bases.
Sclerites in S. maxima ranged in length from 0.06 to
5.0 mm. The largest sclentes in S. maxima were elongated
spindles with complex tubercles (sensu Bayer et al., 1983).
The small sclerites (<0.5 mm long) were rods with vol-
cano-like processes and wart clubs (sensu Bayer et al.,
1983). Wart clubs were proportionately more abundant
in the tips than the bases of S. maxima. The large sclerites
(>0.5 mm long) of S. polydactyla were comprised pri-
marily of spindles with complex tubercles. Most of the
spindles were straight; however, a few were bent. The small
sclerites were either thorn clubs or rods with volcano-like
processes. Sclerites from S. polydactyla ranged in length
from 0.08 to 5.0 mm. In S. sp., the larger sclerites were
comprised of spindles with complex tubercles, thorn stars,
and thorn scales. The spindles were frequently bent and
occasionally bifurcated on one end. The tubercles on the
spindles of S. sp. were smaller than those on 5. maxima
and S. polydactyla. The smaller sclerites were rods and
thorn clubs. Sclerites from 5. sp. ranged in length from
0.08 to 4.6 mm.
Effectiveness of sclerites as feeding deterrents
At natural concentrations, all of the sclerites from the
bases and the tips of all three species of Sinularia were
significantly (P < 0.05) deterrent to fishes at Fingers Reef
(Fig. 8). The addition of sclerites from the tips of Sinularia
spp. reduced feeding by 33-44%, whereas sclerites from
the bases of the colonies reduced feeding by 27-33%.
SINULARIA SP.
100
1
80 •
60
40
100
80
60
2
40 -
20-
ij..^
100 T
I 3
^
80 -IT
^r
60 1
o
I
z
UJ
40 •
u
o
2°lii
UJ
ff
u.
100 f " V
UJ
80 j
K
60 n
UJ
cr
40 •
"•ill
100 f 5
80 -1
60 1 1
40 1
lili.
100 -I
6
in
<
S
60
40
20
III,.
100
80
60
40
20 -
100 •
80 •
60
40 -
20 •
0
ll.. ,
SIZE CLASS (mm) SIZE CLASS (mm)
Figure 5. Relative frequencies and masses of sclerites within ten
length classes from six locations within colonies of Sinularia sp. Samples
were taken along transects from the tips of the colonies (location 1) to
the bases of the colonies (location 6). The values for each size class are
means from five colonies ± 1 SD
236
K. L. VAN ALSTYNE ET AL.
SINULARIA POLYDACTYLA
100-
1
80 -
60 -
40
20 -
100 -
2
80 -
60
40 -
20 -
100 -
t,T
3
80-
d
60 '
:REQUENCY
40
20
100
u^
4
UJ
80
h-
60 •
<
_l
LU
(X
40
20
100 -
,„.
5
80 -
60 '
40 -
Illl..
U5
<
dry weight, none of the sclerites from either bases or tips
were significantly deterrent (Figs. 9, 10). Artificial diets
containing sclerites from the bases of the colonies at 10%
by dry weight were consumed at approximately equal rates
SIZE CLASS (mm)
Figure 6. Relative frequencies ar
length classes from six locations within
Samples were taken along transects from the tips of the colonies (location
1 ) to the bases of the colonies (location 6). The values for each size class
are means from five colonies ± 1 SD.
The ability of Sinularia sclerites to deter feeding was
dependent upon the concentration of sclerites added to
the artificial diet. At the low concentrations of 10% by
SINULARIA MAXIMA
100 -i 2
80 -
60-
100
75
1 100
75
1
=*iu.
50
25
50
25
3
60 -
100
100
2
60 -
75 •
75
::A,
50 -
25
50
25
1
100 1
4
80-
100 -
100
O
3
60-
d 7S
75
^iliin..
o 50 '
§ »
0
50
5? 25
-.i w
,ll.
100 -
80 -
1 00 -
LU
4 < '00-
4
> 75 -
75 -
60 -
t—
< 50-
50
vitals .
UJ
25
ill,
„
6
80 •
100 -
100
5
75 -
75
60 -
40 -
50 -
50
,
:iiiiii*j-
25 -
II,
llll.
100 '
100 -
6
6
SIZE CLASS (mm)
75 -
50 -
25 -
75 '
i ::
i,
ill.
nonowoinoina
isses of sclerites within ten
lies of Sinularia nolvdactvla.
o«-»-f>«f><r»n»»*jtn
o o — —
o o — — f* (st i
SIZE CLASS (mm)
SIZE CLASS (mm)
Figure 7. Relative frequencies and masses of sclerites within ten
length classes from six locations within colonies of Sinularia maxima.
Samples were taken along transects from the tips of the colonies (location
1 ) to the bases of the colonies (location 6). The values for each size class
are means from five colonies ± 1 SD.
STRUCTURAL DEFENSES IN SOFT CORALS
237
100
80
60-
c 40-
o
20-
1.4
Tips Base Tips Base Tips Base
S. maxima S. polydactyla S. sp.
^H Control | | Sclentes Added
Figure 8. Mean amount consumed by generalist fishes at Fingers
Reef of artificial diet with sclerites from Sinn/aria spp. (white bars) and
without sclentes (dark bars). Numbers in parentheses at the bases of the
bars are P values from Wilcoxon Signed Ranks tests for paired comparisons.
10
S. maxima
30 39
[Sclerites] (%d.w.)
• S. polydactyla
SO
S. sp.
Figure 10. Feeding deterrence ofSinularia tip sclerites towards gen-
eralist carnivorous fishes at different sclerite concentrations. Points rep-
resent the mean feeding rate on artificial diet with sclerites relative to
control diet lacking sclerites. Each value represents the results of a single
two-way choice experiment. Circled points denote experiments in which
sclerites were significantly deterrent (P < 0.05).
to artificial diet lacking sclerites (Fig. 9); artificial diets
containing tip sclerites from S. maxima and S. sp. at 10%
by dry weight were consumed at a greater rate than the
control diets (Fig. 10). The relative consumption rate of
artificial diets with sclerites relative to control diets lacking
sclerites decreased with increasing sclerite concentration.
All of the sclerites except those from the bases of 5. sp.
were significantly deterrent (P < 0.05) at concentrations
BASES
20
S. maxima
40 60
[Sclerites] (%d.w.)
S polydactyla
80
100
S. sp.
Figure 9. Feeding deterrence ofSinularia base sclentes towards gen-
eralist carnivorous fishes at different sclente concentrations. Points rep-
resent the mean feeding rate on artificial diet with sclerites relative to
control diet lacking sclerites. Each value represents the results of a single
two-way choice experiment. Circled points denote experiments in which
sclerites were significantly deterrent (P < 0.05).
greater than 25% by dry weight; S. sp. base sclerites were
not significantly deterrent (P > 0.05) at 50% by dry weight,
but were deterrent at 82% by dry weight.
Intracolony differences in sclerite effectiveness
When directly tested in paired feeding experiments,
sclerites from the bases of the colonies were more effective
feeding deterrents than sclerites from the tips of the col-
onies (Fig. 1 1 ). The amount of feeding on artificial diets
Sinulana Sclentes
100
S maxima S polydactyla
•I Tips m Base
S.sp.
Figure 1 1 . Mean amount consumed by generalist carnivorous fishes
of artificial diet containing tip (dark bars) and base (white bars) sclerites
in paired feeding experiments (n = 19 to 20). Numbers in parentheses
are P values from Wilcoxon Signed Ranks tests for paired comparisons.
238
K. L. VAN ALSTYNE ET AL.
containing sclerites from the bases of the colonies was
18% (in S- sp.) to 51% (in S. maxima) lower than the
consumption of diets containing tip sclerites.
Because both the morphologies and concentrations of
sclerites differ between the bases and tips of the colonies,
differences in the deterrence of tip and base sclerites could
result from differences in the sizes and shapes of the scler-
ites or from differences in their amounts. To sort out the
relative contributions of sclerite morphology and concen-
tration to predator deterrence, experiments were con-
ducted in which fishes were offered four choices of foods
simultaneously: diet containing tip sclerites at tip con-
centrations, diet containing tip sclerites at base concen-
trations, diet containing base sclerites at tip concentra-
tions, and diet containing base sclerites at base concen-
trations. These experiments demonstrated that the ability
of sclerites to deter feeding was dependent upon both the
morphology of the sclerites and their concentrations;
however, concentration is a more important factor than
morphology in Sinularia spp. (Table III). Feeding rates
were higher on artificial diets containing sclerites at the
lower tip concentrations than the higher base concentra-
tions, regardless of sclerite type (Fig. 12). In S. maxima,
the smaller tip sclerites were less effective deterrents than
the larger base sclerites at high concentrations, but were
only slightly less effective at low concentrations. In S.
polydactyla, tip sclerites were more effective deterrents
than base sclerites at both high and low concentrations.
In S. sp., tip and base sclerites were equally effective at
low concentrations; however, at high concentrations, tip
sclerites were less effective deterrents.
Discussion
This and other studies have clearly demonstrated the
potential for a defensive role for alcyonarian sclerites. We
have shown that sclerites from three species of Sinularia
in Guam, when incorporated into an artificial diet, re-
duced feeding by generalist carnivorous fishes in the field
(Fig. 4). Similar studies with sclerites from the Caribbean
gorgonians Pseudoptcrogorgia acerosa (Harvell et at..
100
80
E 60
40
20
Sclerites
I
1
•
S. maxima S. polydactyla
S. sp.
<jj§ Tips. Low | | Tips, High [ | Base, Low jggj Base, High
Figure 12. Results of four-way feeding experiments. Vertical bars
represent mean feeding rates on tip or base sclerites at natural tip (low)
or natural base (high) concentrations. Vertical lines represent ± 1 SE.
Statistical analyses are presented in Table III.
1988) and Gorgonia ventalina (Van Alstyne and Paul, in
press) have also demonstrated that sclerites can deter
feeding in the field. However, not all alcyonarian sclerites
are deterrent at natural concentrations. For example,
sclerites from the white whip, Jituceela sp., naturally occur
at concentrations of ~45% by dry weight, a similar con-
centration to those found in Sinularia: but, Junceela
sclerites are not deterrent towards carnivorous fishes at
natural dry weight concentrations (Paul and Van Alstyne,
in prep.).
Invertebrate spicules and sclerites vary widely in size,
shape, and concentration. We have demonstrated that, in
these three species of Sinularia, concentration is more
important than morphology in determining the ability of
sclerites to deter the generalist predators of Guam. Sin-
ularia sclerites were increasingly effective as deterrents
until concentrations of about 30-50% by dry weight, at
which point deterrence leveled off (Figs. 9, 10). The level
at which maximum effectiveness was reached was deter-
mined by the sizes and shapes of the sclerites. The differ-
Table III
Results of'hw-mn- Kruskal-Wallis tests on data from four choice feeding experiments
Location of sclerites
Concentration of sclerites
Species
Data are presented in Figure 12.
Location* concentration
Sinularia maxima
2.696
0.096
29.13
0.000
1.265
0.260
Sinn/aria polydactyla
7.057
0.008
9.136
0.003
0.071
0.786
Sinn/aria sp.
1.972
0.156
17.39
0.000
1.238
0.265
STRUCTURAL DEFENSES IN SOFT CORALS
239
ences in deterrence of sclerites from the tips and bases of
the colonies (Fig. 1 1 ) was primarily a result of concentra-
tion differences of sclerites in these two regions (Table
III), not of morphological differences between tip and base
sclerites.
Although differences in the sizes and shapes of sclerites
between the tips and the bases of the colonies had a lesser
impact on the relative effectiveness of structural defenses
than sclerite concentration, sclerite morphology is still an
important determinant of feeding deterrence. Differences
in sclerite morphologies between the tips and bases of 5.
polydactyla colonies significantly affected feeding by fishes
off Guam (Table III). Two components of sclerite mor-
phology may influence their effectiveness as feeding de-
terrents: size and shape. The effects of sclerite morphology
may be more important when making interspecific com-
parisons of deterrence. For example, sclerites from
Junceela sp. do not deter feeding by fishes at Fingers Reef
(Paul and Van Alstyne, in prep.). These Junceela sclerites
are similar in size and concentration to many of the
smaller Sinularia sclerites, but, they differ in shape (Paul
and Van Alstyne, in prep.). Sclerite shape probably has a
significant effect on function, particularly in determining
the ability of the sclerite to deter potential predators.
Although our experiments have demonstrated that
Sinularia sclerites can deter feeding by fishes, predator
deterrence may not be the primary function of these
structures. Spicules from marine invertebrates also play
a role in structural support for colonies, increasing the
stiffness of connective tissues by acting like reinforcing
fibers (Koehl, 1 982). Small sclerites tend to increase colony
stiffness more than large sclerites; however, stiffness, like
predator deterrence, increases with increasing sclerite
concentration (Koehl, 1982). Further work is needed to
clarify the structural and defensive functions of inverte-
brate spicules.
The differences in sclerite morphologies and concen-
trations within Sinularia colonies may reflect differences
in the function of sclerites in different locations within
the colony or differences in constraints upon sclerite use.
Sclerites are most concentrated in the bases of the colonies
(Fig. 4), making the bases less susceptible to attack than
the tips (Fig. 12, 13, Table III). However, it is the tips of
the colonies that receive the majority of attacks by pred-
ators (Wylie and Paul, 1989). The lack of high levels of
sclerites in the colony tips may reflect constraints on
sclerite use. The presence of large quantities of sclerites
may interfere with the functioning of the soft coral polyps,
which are found only in the tips of the colonies. Alter-
natively, sclerites in the bases of the colonies may serve
primarily as structural support rather than predator de-
terrents. The presence of high concentrations of predator-
deterrent extracts in the tips of the colonies (Wylie and
Paul, 1989; Van Alstyne et at., in prep.) supports the hy-
pothesis that sclerite concentrations in the tips of colonies
are under functional constraints.
Acknowledgments
The authors are grateful to D. Carandang-Liberty, K.
Foster, K. Kuetzing, H. Sanger, and K. Sonada for the
many hours spent attaching o-rings onto safety pins. We
are also grateful to these individuals and the attendees of
the UOG POETS club for spending many more hours
taking the o-rings off safety pins. We are indebted to D.
Carandang-Liberty, K. Foster, B. Irish, L. Meyer, H. San-
ger, and K. Sonada for their assistance with the field ex-
periments. We also thank S. Murray of the California
State University at Fullerton for use of his image analysis
system, J. Smith of CSUF for advice on statistical analyses,
and A.O.D. Willows and the staff of the Friday Harbor
Laboratories for the use of the scanning electron micros-
copy facilities. This manuscript greatly benefitted from
the comments of two anonymous reviewers. This research
was funded by NIH grant GM 38624 to VJP and a Guyer
postdoctoral fellowship from the University of Wisconsin
to KLV. This is contribution number 31 1 of the Univer-
sity of Guam Marine Laboratory.
Literature Cited
Bayer, F. M. 1956. Octocorallia. Pp. F166-F231 in Treatise on In-
vertebrate Paleontology, Pan F, Coelenterata, R. C. Moore, ed. Uni-
versity of Kansas Press, Lawrence, KS.
Bayer, F. M. 1961 . The shallow water Octocorallia of the West Indian
region. Studies on the Fauna of Curacao and Other Caribbean Islands
12: 1-373.
Bayer, F. M., M. Grasshoff, and J. Verseveldt. 1983. Illustrated Tri-
lingual Glossary of Morphological and Anatomical Terms Applied to
Octocorallia. E. J. Brill. Leiden. 75 pp.
Coll, J. C., D. M. Tapiolas, B. F. Bowden, L. Webb, and H. Marsh.
1983. Transformation of soft coral (Coelenterata: Octocorallia)
terpenes by Ovula ovum (Mollusca: Prosobranchia). Mar Biol. 74:
35-40.
Gawel, M. J. 1977. The common shallow-water soft corals (Al-
cyonacea) of Guam. M.Sc. Thesis, University of Guam. Agana.
Guam. 201 pp.
Harvell, C. D., and W. Fenical. 1989. Chemical and structural
defenses of Caribbean gorgonians (Pseudopterogorgia spp.): in-
tracolony localization of defense. Limnol. Oceanogr. 34: 380-
387.
Harvell, C. D., W. Fenical, and C. H. Greene. 1988. Chemical and
structural defenses of Caribbean gorgonians (Pseudopterogorgia spp.)
I: Development of an //; situ feeding assay. Mar Ecol. Prog. Ser. 49:
287-294.
Harvell, C. D., and T. H. Suchanek. 1987. Partial predation on tropical
gorgonians by Cyphoma gibbosum (Gastropoda). Mar. Ecol. Prog.
Ser 38: 37-44.
Kingsley, R. J. 1984. Spicule formation in the invertebrates with special
reference to the gorgonian Leptogorgia virgiilata. Am. Zool. 24: 883-
891.
240
K. L. VAN ALSTYNE ET AL.
Kingsley, R. J., M. Tsuzaki, N. Watabe, and G. L. Mechanic.
1990. Collagen in the spicule organic matrix of the gorgonian Lep-
togorgia virgiilat/t. Biol. Bull. 179: 207-213.
Koehl, M. A. R. J982. Mechanical design of spicule-reinforced con-
nective tissue. J. Exp. Biol 98: 239-268.
LaBarre, S., J. C. Coll. and P. W. Sammarco. 1 986. Defensive strategies
of soft corals (Coelenterata: Octocorallia) of the Great Barrier Reef.
II: The relationship between toxicity and feeding deterrence. Biol.
Bull. 171: 625-636.
Paul, V. J., and K. L. Van Alstyne. 1987. Chemical defense and chem-
ical variation in some tropical Pacific species of Halimeda (Hali-
medaceae; Chlorophyta). Coral Reefs 6: 263-269.
Pawlik, J. R., M. T. Burch, and W. Fenical. 1987. Patterns of chemical
defense among Caribbean gorgonian corals: a preliminary survey. J.
Exp. Mar. Biol. Ecol. 108: 55-66.
Sammarco, P. W., S. LaBarre, and J. C. Coll. 1987. Defensive strategies
of soft corals (Coelenterata: Octocorallia) of the Great Barrier Reef:
III. The relationship between icthyotoxicity and morphology. Oec-
0/ogza74:93-101.
Tursch, B., and A. Tursch. 1982. The soft coral community on a shel-
tered reef quadrat at Laing Island (Papua New Guinea). Mar. Biol.
68:321-332.
Van Alstyne, K. L., and V. J. Paul. 1992. Chemical and structural
antipredator deterrents in the sea fan Gorgonia ventalina: effects
against generalist and specialist predators. Coral Reefs In
press.
Verseveldt, J. 1977. Octocorallia from various localities in the Pacific
Ocean. Zoo/. Verhand. Leiden 150: 1-52.
Verseveldt, J. 1980. A revision of the genus Sinularia May (Octocor-
allia, Alcyonacea). Zoo/. Verh. Leiden 178: 1-165.
Wylie, C. R. 1988. Chemical defenses in three species of Sinularia
(Coelenterata, Alcyonacea): effects against the predator Cheatodon
unimaculalus (Perciformes). M.Sc. Thesis, University of Guam,
Mangilao, Guam. 38 pp.
Wylie, C. R., and V. J. Paul. 1989. Chemical defenses in three species
of Sinularia (Coelenterata, Alcyonacea): effects against generalist
predators and the butterflyfish Chaetodon unimaculalus Bloch. J.
Exp. Mar. Biol. Ecol. 129: 141-160.
Reference: Biol. Bull 182: 241-247. (April. 1992)
Characterization of Two Novel Neuropeptides From
the Sea Cucumber Holothuria glaberrima
LUCY DIAZ-MIRANDA1. DAVID A. PRICE2. MICHAEL J. GREENBERG2,
TERRY D. LEE3, KAREN E. DOBLE2, AND JOSE E. GARCIA-ARRARAS1
^Department of Biology, University of Puerto Rico, Rio Piedras, Puerto Rico 0093 J; 2The Whitney
Laboratory, St. Augustine, Florida 32086; and ^Division of Immunology, Beckman Research
Institute of the City of Hope, Duarte, California 91010
Abstract. Two peptides were purified from intestinal
extracts of a sea cucumber, Holothuria glaberrima, by
high pressure liquid chromatography (HPLC). The pep-
tides were detected by a radioimmunoassay (RIA) based
on an antiserum raised to the molluscan peptide, pGlu-
Asp-Pro-Phe-Leu-Arg-Phe-NH2 (pQDPFLRFamide).
Automated sequencing and mass spectrometry indicate
that the isolated peptides are: Gly-Phe-Ser-Lys-Leu-Tyr-
Phe-NH2 (GFSKLYFamide) and Ser-Gly-Tyr-Ser-Val-
Leu-Tyr-Phe-NH2 (SGYSVLYFamide). These are the first
peptides to have been isolated from a member of the echi-
noderm class Holothuroidea.
The antiserum used in the RIA of the peptides was also
employed in localizing immunoreactive nerve cells and
fibers in the intestine of//, glaberrima. The immunohis-
tochemical results suggest that these peptides might be
responsible for the FMRFamide-like immunoreactivity
reported earlier. Sequence similarities between
GFSKLYFamide, SGYSVLYFamide, and a pair of pep-
tides previously isolated from starfish lead us to propose
that all four molecules are members of a family of peptides
acting as neurotransmitters in echinoderms.
Introduction
Very few echinoderm neuropeptides have been char-
acterized. For example, the sequence of the first neuro-
peptide detected in this phylum — i.e., gonad-stimulating
substance (GSS) from starfish (Chaet and McConnaughy,
1959)— is still unknown (references in Cobb, 1988). Re-
cently, FMRFamide-like immunoreactivity was detected
in the nervous system of the starfish Asterias rubens, and
Received 10 October 1991; accepted 27 January 1992.
immunoreactive nerve fibers were found in the area of
the tube feet, suggesting that FMRFamide might be reg-
ulating the process of locomotion (Elphick el al, 1989).
Subsequently, two novel neuropeptides from the starfish
A. rubens and A. forbesi were identified: GFNSALM-
Famide and SGPYSFNSGLTFamide, and the previously
reported FMRFamide-like immunoreactivity in A. rubens
was attributed to these peptides (Elphick et al.. 1991).
Peptide immunoreactivity has also been demonstrated
in members of another echinoderm class, the Holothu-
roidea. For example, cholecystokinin (CCK)-like immu-
noreactivity occurs in neurons and in a plexus of fibers
in the intestines of Holothuria mexicana, Holothuria gla-
berrima, and Stichopus badionotus (Garcia-Arraras et al.,
199 la). Similarly, FMRFamide-like immunoreactivity
was reported in cells and fibers of the intestine of H. gla-
berrima (Garcia-Arraras et al., 1991b). The location of
these immunoreactive substances suggests that they have
a role in the regulation of digestive physiology, and indeed
peptides of the CCK family do induce a partial relaxation
of the intestinal musculature (Garcia-Arraras et al..
199 la). These results notwithstanding, none of the en-
dogenous peptides in the nervous system of sea cucumbers
had been identified before the present study was under-
taken.
In this report, we describe the isolation and character-
ization of two peptides from the digestive system of H.
glaberrima. In addition, we provide histochemical evi-
dence for the presence of these peptides in the enteric
nervous system of the sea cucumber.
Materials and Methods
Specimens of//, glaberrima ( 10- 1 5 cm in length) were
collected from the rocky intertidal zone of the north coast
241
242
L. DIAZ-MIRANDA ET AL
of Puerto Rico. The animals were either used immediately
or, in some cases, maintained in marine aquaria at the
Department of Biology of the University of Puerto Rico.
Digestive system extracts
Four extracts of sea cucumber digestive systems were
prepared as follows. First, 1 3 to 25 animals were sectioned
with a razor blade, just posterior to the calcareous ring.
The body wall was slit open, and the intestinal tract, in-
cluding the esophagus and the small and large intestine,
together with adherent pieces of hemal vessel and respi-
ratory tree, were removed. The tissue (37-113 g wet
weight) was placed in a fourfold excess of acetone and left
at — 20°C for 48 h. The supernatant was then filtered
through Whatman #1 paper, and the acetone and part of
the water were removed on a rotary evaporator. The
aqueous portion was acidified to 0.02 M acetic acid, cen-
trifuged at 2500 X g. and the supernatant dried in a Speed-
Vac (Savant). The dried sample was reconstituted in
aqueous 0. 1% trifluoroacetic acid (TFA), centrifuged, and
filtered.
Purification
The reconstituted sample was pumped onto a Brownlee
C8 reverse phase column (Aquapore RP300: 220 X 4.6
mm or Prep 10 Aquapore Octyl 100 X 10 mm), according
to the procedure described by Price el al. ( 1990a). Once
loaded, the column was rinsed with aqueous solvent (0. 1 %
TFA) until the UV absorbance fell to near baseline. The
eluting solvent was rapidly changed (by a step or 1 min
gradient) to 20 or 30% of organic solvent (80% acetonitrile
containing 0.1% TFA), whereupon a linear gradient was
started with a 1%/min increase in the organic solvent up
to 50 or 60% organic. Half-minute fractions were col-
lected, and 2 n\ aliquots were taken from each fraction
for the RIA.
Further purification was also done on Brownlee C8 RP-
300 columns (220 X 2.1 mm or 220 X 4.6 mm). The
columns were developed, either with TFA/acetonitrile
gradients as above, or with aqueous 0.05% heptafluoro-
butyric acid (HFBA) and 80% acetonitrile containing
0.05% HFBA.
Radioimmunoassay (RIA)
A rabbit antiserum (Q2), raised against pQDP-
FLRFamide coupled to thyroglobulin (Price et al.. 1990b),
was diluted 1:500 for use in the RIA. lodinated
pQYPFLRFamide served as the tracer.
Sequencing and spectrometry
The most immunoreactive fraction within each pure
peak was analyzed. The fraction was divided in half: one
half was dried in a Speed- Vac for FABms, and the other
half (about .1 ml) was applied (in 3 portions with inter-
mediate drying) directly to a pre-conditioned glass-fiber
filter disk containing 3 mg of Polybrene. The disk was
placed in the sequencer (Applied Biosystems 470A gas-
phase sequencer with an on-line 1 20a PTH analyzer), and
the PTH-amino acid derivatives in each cycle were iden-
tified by their retention times and quantitated by com-
parison of the peak areas to standards (performed by B.
Parten, University of Florida Interdisciplinary Center for
Biotechnology Research, Protein Sequencing Core Facil-
ity, Gainesville). The FABms analysis was carried out on
a JEOL HX100HF magnetic sector mass spectrometer,
as described in Bulloch et al. (1988).
Synthetic peptide
The peptides GFSKLYFamide and SGYSVLYFamide
were synthesized on an Applied Biosystems synthesizer
by the Protein Chemistry Laboratory of the University of
Florida's Interdisciplinary Center for Biotechnology Re-
search; t-Boc protecting groups were used. The peptides
were deprotected and removed from the resin with tri-
fluoromethanesulfonic acid (Applied Biosystems proto-
col), purified by HPLC, and quantified by amino acid
analysis (Hitachi 835 analyzer).
Immunohistochemistry
For the histochemical study, the procedure described
by Garcia- Arraras et al. (199 la) was followed. In brief,
isolated portions of the large and small intestines of H.
glaberrima were fixed in picric acid-formaldehyde mixture
overnight at 4°C. The tissue sections (10-12 ^m) were
treated with antiserum Q2 (1:500) or with an antiserum
against FMRFamide (#8 3i 2s; 1:500) (Garcia-Arraras et
a/., 1991b). As a control, theQ2 antiserum was incubated
with 10 Mg/ml of GFSKLYFamide, FMRFamide (Pen-
insula), or FLRFamide (Sigma) for 24 h before being ap-
plied to the tissue sections.
Results
Fractionation oj extracts
Each of the four gut extracts was fractionated with a
somewhat different series of HPLC steps. We discovered,
finally, that the simplest way to purify the immunoreactive
peptides was to select the immunoreactive fractions after
each step, and to run them back through the same column
under the same conditions (Fig. 1). This finding was cer-
tainly not expected. The method works because the extract
behaves anomalously; i.e., the immunoreactive peaks shift
to an ever earlier elution time during each successive step
of purification (e.g.. compare Figs. 1C2 and 1C3). More-
over, even the order of elution of the immunoreactive
NOVEL NEUROPEPTIDES FROM SEA CUCUMBERS
243
B4 100-1
r 2
10 20
Elution Time (min)
10 20
Elution Time (min)
244
L. DIAZ-MIRANDA ET AL
peaks changes (Fig. 1). This relative shift in elution po-
sition was clearly observed in all but the first of the four
extracts.
One peptide, already purified from the first extract, was
readily and unambiguously identified as GFSKLYFamide
when we obtained its molecular ion (860.4) and sequence
(Fig. 2a). The calculated value of the amide of this se-
quence is 860.5, whereas that of the free acid is 861.5.
Thus the molecular ion confirmed the presence of the C-
terminal amide, which had been inferred from the im-
munoreactivity; in contrast, the PTH derivatives of phe-
nylalanine and its amide are indistinguishable in normal
Edman sequencing. The first extract contained other mi-
nor immunoreactive peaks; one of these was analyzed by
FABms and sequenced; this product, found again in the
fourth extract, will be described further below.
From the second extract, we obtained a molecular ion
(934.6) and the partial sequence of a second peptide,
SGXSVLXFamide (where X could be either tyrosine or
methionine sulfone). The first peptide (GFSKLYFamide)
did not appear in this extract.
The third extract contained two immunoreactive peaks.
Using mass spectrometry, we again identified GFSKLY-
Famide (860 molecular ion) and SGXSVLXFamide (934
molecular ion).
The fractionation of the fourth extract was undertaken
to resolve the ambiguities left after the first three, and the
HPLC runs leading to the purification of the two most
immunoreactive peaks are shown in Figure 1.
The first HPLC step in this last purification yielded a
broad, irregular peak of immunoreactivity (Fig. 1 A). From
the earlier half of this, we succeeded, after four steps, in
purifying a peak (at 15 min in Fig. 1B4) that sequenced
as SGYSVLYF (Fig. 2B). The calculated molecular ion
for this peptide — with its C-terminal amidated — is 934.5,
and this is in good agreement with the molecular ion
(934.6) found earlier. The small immunoreactive peak at
1 6 min in Figure 1 B4 had the same sequence as the main
peak, so it is probably just a tail of the main peak.
From the later half of the initial broad immunoreactive
peak obtained in the first step of the purification (Fig.
1A), two peaks were resolved in the second (Fig. 1C1).
The major (and earlier-eluting) of these peaks co-eluted
with synthetic GFSKLYFamide.
The second peak, eluting at 23.5-25 min, was pooled
with a similar small peak that had eluted a few minutes
456
Cycle Number
10
Figure 2. Amino acid sequences of the purified peptides. The yields
of the pertinent PTH amino acid derivatives at each cycle are plotted
and the amino acid assigned to each position is shown. (A) The peptide
having an 860 molecular ion. (B) The peptide having a 934 molecular
ion.
after SGYSVLYFamide (see arrows in Figs. 1B1 and 1C1).
After purification (not shown), these pooled minor peaks
yielded the sequence GFSXLYF, which corresponds to
that of the synthetic peptide, except that no lysine (or
other PTH derivative) appeared in cycle 4. This peak had
a molecular ion of 958, which is 98 larger than that ex-
pected for the lysine containing amidated peptide. A peak
with the same relative elution time, and a similar molec-
Figure 1. HPLC fractionation of a sea cucumber gut extract. The ultraviolet absorbance at 210 nm
(solid line) and the immunoreactivity (histogram) are shown for each HPLC run. The initial fractionation
(A) was done on a PreplO Octyl column (10 x 100 mm; 4 ml/mm) with a 30 min gradient from 16 to 40%
acetonitrile in water with 0.1% trifluoroacetic acid. Subsequent runs (B 1-4: Cl-3) were done on an RP-300
column (2.1 > 220 mm; 0.5 ml/min) with the same gradient. The arrows in Bl and Cl indicate peaks that
were pooled and subsequently purified (not shown; see text). The full scale absorbance in the top trace is
2.0 and is 0.5 AU in all the subsequent traces.
NOVEL NEUROPEPTIDES FROM SEA CUCUMBERS
245
ular ion (958.6) and sequence, had also appeared in the
first extract, as briefly noted above.
We suspect that the peptide GFSXLYFamide has a
derivatized lysine side chain. For example, a peptide with
the lysine amino group amidated by hexanoic acid would
give such a molecular ion. and such a derivative could
occur naturally in the sea cucumber. It is more likely,
however, that the derivative is a substituted oxazolidine
ring, formed by the condensation, with the lysine, of two
molecules of acetone (58 + 58) with the loss of a water
(—18). This derivatization would add 98 to the molecular
weight of the peptide. In summary, this peptide may very
well be an artifact of the purification.
Synthetic peptides
The synthetic peptides GFSKLYFamide and SGYSV-
LYFamide were purified by HPLC after deprotection.
Each peptide co-eluted with its natural counterpart on
HPLC. An amino acid analysis of GFSKLYFamide gave
the following composition: Gly 1.05, Phe 1.75, Ser 1.0,
Lys 1.1, Leu 1.25, and Tyr 1.05. The composition of
SGYSVLYFamide was: Gly 1.00, Phe .90, Ser 2.03, Leu
1.11, Tyr 1.99, Val 1.33.
Q2 immunoreactivity
The tissue distribution of immunoreactivity to Q2 (the
antiserum used in the characterization of GFSKLYFam-
ide and SGYSVLYFamide) was determined by immu-
nohistochemistry. Results with antibodies Q2 and #8 (the
latter an antiserum against FMRFamide) were similar.
Cells and fibers located in the outer connective tissue (se-
rosa) of the small and large intestine were labeled, as were
single fiber-like projections in the submucosal layer (Fig.
3). In addition, a strong nerve plexus was observed in the
mesentery next to the muscular layer. This nerve plexus
was continuous with the serosal nerve plexus. In contrast
to the FMRFamide-like immunoreactivity (Garcia-Arra-
ras et ai, 1991b), Q2 also labeled a group of cells located
in the submucosal layer of the intestines, similar to what
has been described as morula cells (Hetzel, 1963, 1965),
but this reaction does not seem to be specific.
When the Q2 antiserum was preabsorbed with
GFSKLYFamide (12 jiM), no immunoreactivity to Q2
was observed in the cells and fibers of the mesentery, se-
rosa, or submucosa; but the morula cells continued to be
labeled. These results suggest that other antibodies in the
Q2 serum recognize unrelated substances. The peptides
FMRFamide ( 1 7 fiM) and FLRFamide ( 1 7 pM) were also
used for preabsorption of Q2, but they could not com-
pletely block the observed immunoreactivity of the cells
and nerve fibers in the serosa, or of the nerve fibers in the
submucosa. The Q2 immunoreactivity of the morula cells
Figure 3. Transverse sections of the large intestine of//, glaberrima
showing Q2-like immunoreactivity. A. Immunoreactive nerve fibers at
the level of the serosa and longitudinal muscle. Most of these fibers were
associated with the longitudinal muscle (arrowhead). B. Arrowhead points
at one immunoreactive fiber extending throughout the submucosa layer.
Asterisk: endogenous fluorescence. Magnification: A. X405; B. X270.
was not blocked, whether the antiserum was preabsorbed
with FMRFamide or with FLRFamide.
Discussion
We have purified two peptides from gut extracts of the
sea cucumber H. glaberrima, using high pressure liquid
chromatography for separation and radioimmunoassay
for detection. These peptides — GFSKLYFamide and
SGYSVLYFamide — were completely characterized by
microsequencing and mass spectrometry, and are the first
to have been isolated from the echinoderm class Holo-
thuroidea.
Two related peptides were isolated earlier from another
echinoderm class, the Asteroidea: i.e.. GFNSALMFamide
and SGPYSFNSGLTFamide from the starfishes Asierias
forbesi and A. ntbens (Table I; Elphick et al., 1991). These
peptides, like those of the sea cucumber reported here,
were detected on the basis of their binding to an antiserum,
Q2, raised against pQDPFLRFamide (Price et a/.. 1990b).
In both studies, furthermore, the Q2 antiserum was orig-
inally selected because the aim was to characterize putative
FMRFamide-related peptides (FaRPs) that had been
246 L. DIAZ-MIRANDA ET AL.
Table I
Amino acid sequences of SxLxFamide1 peptides isolated from Echinodermata
CLASS
Species
Sequence
Ref.
HOLOTHUROIDEA
Holothuria glaberrima
ASTEROIDEA
Asterias forbesi
A. rubens
Gly-Phe-SER-Lys-LEU-Tyr-PHE-NH2
Ser-Gly-Tyr-SER-Val-LEU-Tyr-PHE-NH2
Gly-Phe-Asn-SER-Ala-LEU-Met-PHE-NH2
Ser-Gly-Pro-Tyr-Ser-Phe-Asn-SER-Gly-LEU-Thr-PHE-NH2
1 Ser-x-Leu-x-Phe-NH: . where the positions "x" can be occupied by any residue.
• This report.
3 Elphick ei al .. 1991.
identified by immunocytochemistry ( Elphick el al, 1989;
Garcia- Arraras et a/., 1991b).
FaRPs, defined liberally, have now been isolated from
many animal groups, including coelenterates, nematodes,
annelids, arthropods, and even vertebrates (reviewed by
Price and Greenberg, 1989; Greenberg and Price, 1992),
and a penultimate arginyl residue is not only a common
feature of this extended family, but has been shown to be
critical for physiological activity (e.g., see Kobayashi and
Muneoka, 1986). Antiserum Q2 should have identified
most FaRPs that might have been present in our extracts,
but in fact, not one of the four neuropeptides sequenced
from echinoderms has the penultimate arginine. Our re-
sults, therefore, when taken together with the evidence
obtained from the starfish (Elphick et al., 1991), suggest
that authentic FaRPs are absent from the Echinodermata.
If the above assertion is correct, we must account for
the FMRFamide-like immunoreactivity described in sea
cucumbers (Garcia-Arraras el al., 199 Ib) and in starfish
(Elphick el al.. 1989). In the sea cucumber H. glaberrima,
immunoreactivity to FMRFamide has been reported in
the radial nerves, in nerve plexuses of the esophagus, and
in the enteric nervous system (Garcia-Arraras et al.,
1991b). The distribution of immunoreactivity observed
with the Q2 antibody in the gut of //. glaberrima is similar,
if not identical, to that observed with antibodies against
FMRFamide. The localization of Q2-binding to nerve
cells and fibers in the sea cucumber intestine suggests that
the peptides recognized by this antibody occur in the en-
teric nervous system of holothurians, and that they may
be involved in neural transmission or modulation.
The story in the starfish Asterias rubens is similar: i.e.,
FMRFamide-like immunoreactivity was detected in the
radial nerve cords, the circumoral nerve ring, and the sub-
epithelial nerve plexus of the tube feet (Elphick et al.,
1989), and this immunoreactivity has been attributed to
the two isolated peptides from the starfishes by Elphick
et al. (1991). In conclusion, the FMRFamide-like im-
munoreactivity previously reported in both echinoderm
classes might be due to the presence of the isolated peptides
reacting with the anti-FMRFamide serum.
As Table I illustrates, the two sea cucumber peptides
have five of seven residues in common, and the two star-
fish peptides have five of eight identical; but the most
similar sea cucumber and starfish peptides share only four
of eight residues. The clear unifying feature of these four
echinoderm peptides is the C-terminal sequence SxLx-
Famide, where the positions indicated by "x" can be oc-
cupied by any other residue. We therefore propose that
this periodic sequence of serine, leucine, and phenylala-
nine defines a novel family of peptides present in the
Echinodermata.
Acknowledgments
This research was supported by funds from the Re-
source Center for Sciences and Engineering, the Patricia
Robert Harris Fellowship to LDM, the Grass Foundation
(to the Whitney Lab), the National Science Foundation
(BNS-8801538 to JGA), and the National Institute of
Health (HL28440 to MJG). We would like to thank the
Protein Chemistry Core Facility of the University of
Florida for peptide sequencing and peptide synthesis, and
we also acknowledge the help of the Interdisciplinary
Center for Biotechnology Research (ICBR) of the Uni-
versity of Florida. We are very grateful to Mr. Dietmar
Nieves and Ms. Lynn Milstead for their excellent assis-
tance with the preparation of figures.
Literature Cited
Bulloch, A. G. M., D. A. Price, A. D. Murphy, T. D. Lee, and H. N.
Bowes. 1988. FMRFamide peptides in Helisoma: identification and
physiological actions at a peripheral synapse. / Neurosci. 8: 3459-
3469.
Chaet, A. B., and R. A. McConnaughy. 1959. Physiologic activity of
nerve extracts. Biol Bull 117: 407-408.
NOVEL NEUROPEPTIDES FROM SEA CUCUMBERS
247
Cobb, J. L. S. 1988. Neurohumors and neurosecretion in echinoderms:
a review. Comp. Biochem. Physiol. 91C: 151-158.
Elphick, M. R., D. A. Price, T. D. I,ee, and M. C. Thornd> ke. 1991 . The
SALMFamides: a new family of neuropeptides isolated from an echi-
noderm. Proc. Roy. Soc. B 243: 121-127.
Elphick, M. R., R. H. Emson, and M. C. Thorndyke. 1989.
FMRFamide-like immunoreactivity in the nervous system of the
starfish Aslerias rubens. Biol. Bull. 177: 141-145.
Garcia-Arraras, J. E., I. Torres-Avillan, and S. Ortiz-Miranda.
1991a. Cells in the intestinal system of holothurians(Echinoderm-
ata) express cholecystokinm-like immunoreactivity. Gen. Comp. En-
docrinol. 83: 233-242.
Garcia-Arraras, J. E., I. Enamorado-Ayala, I. Torres-Avillan, and V '.
Rivera. 1991b. FMRFamide-like immunoreactivity in cells and fi-
bers of the holothunan nervous system. Neurosci. Letters 132: 199-
202.
Greenberg, M. J., and D. A. Price. 1992. Relationships among the
FMRFamide-like peptides. Prog Brain Res. (in press).
Hetzel, 11. R. 1963. Studies on holothunan coelomocytes. I. A survey
ofcoelomocyte types. Biol. Bull. 125: 289-301.
Hetzel, H. R. 1965. Studies on holothunan coelomocytes. II. The origin
of coelomocytes and the formation of brown bodies. Biol. Bull 128:
102-112.
Kobayashi, M., and Y. Muneoka. 1986. Structural requirements for
FMRFamide-like activity on the heart of the prosobranch Rapana
thomasiana. Comp. Biochem. Physiol. 84C: 349-352.
Price, D. A., and M. J. Greenberg. 1989. The hunting of the FaRPs:
the distribution of FMRFamide-related peptides. Biol. Bull. 177: 198-
205.
Price, D. A., K. E. Doble, T. D. Lee, S. M. Galli, B. M. Dunn, B. Parten,
and D. II. Evans. 1990a. The sequencing, synthesis, and biological
actions of an ANP-like peptide isolated from the brain of the killifish
Fnndulus heteroclilus. Biol. Bull. 178: 279-285.
Price, D. A., W. Lesser, T. D. Lee, K. E. Doble, and M. J. Greenberg.
i990b. Seven FMRFamide-related and two SCP-related cardioac-
tive peptides from Helix. J. Exp. Biol. 154: 421-437.
Reference: Biol. Bull. 182: 248-256. (April, 1992)
Giant Axons and Escape Swimming in Euplokamis
dunlapae (Ctenophora: Cydippida)
G. O. MACKIE1, C. E. MILLS2, AND C. L. SINGLA1
1 Biology Department, University of Victoria, Victoria, British Columbia. V8W 2Y2, Canada, and
'Friday Harbor Laboratories, University of Washington, Friday Harbor. Washington 98250
Abstract. Euplokamis dunlapae responds to anterior
stimulation by reversing the beat direction of its comb
plate cilia and swimming rapidly backwards. It responds
to posterior stimulation by swimming forwards at an ac-
celerated rate. Video playback and laser monitoring were
used to analyze changes in the pattern of ciliary beating,
while electrical activity was recorded extracellularly. Es-
cape responses occur with latencies of less than 1 50 ms
and involve greatly increased ciliary beat frequencies.
Giant axons run longitudinally along each of the eight
comb rows, as shown by optical and electron microscopy.
They form chains of overlapping neurons, with diameters
of about 12 ^m in life, and conducting at over 50 cm •
s~" as recorded with an extracellular electrode placed di-
rectly over the chain. The giant neurons are synaptically
linked with smaller neurites of the general ectodermal
nerve plexus, with each other, and with the ciliated cells
of the comb plates. They appear to constitute a single
system mediating rapid conduction of signals in either
direction, but a full analysis was not attempted for lack
of sufficient material. Electro-physiological examination
of two other ctenophores (Pleurobrachia and Beroe) gives
no indication of rapid conduction pathways, and these
forms probably lack giant axons.
Introduction
Several cydippid and lobate ctenophores have the ability
to reverse the direction of the power stroke of their comb
plate cilia. In the best-studied example, Pleurobrachia
(Tamm and Moss, 1985; Moss, 1986; Moss and Tamm,
1986, 1987), reversals occur unilaterally as part of feeding
behavior, and make the animal rotate on its axis. Reversals
Received 3 January 1991; accepted 12 December 1991.
can also occur simultaneously in all the comb rows, caus-
ing the animal to swim backwards. Several cases of reverse
swimming have been reported (reviewed by Tamm, 1982).
but none have been studied in detail. We here describe
the responses of an unusual and interesting cydippid
ctenophore, Euplokamis dunlapae, that responds rapidly
to stimulation by forward and reverse swimming. A novel
feature of these responses is that they appear to be me-
diated in part by giant axons that run under the comb
rows.
A taxonomic account of Euplokamis dunlapae (Fig. 1 ).
is given by Mills (1987). The species is probably the most
abundant midwater ctenophore in the Strait of Georgia
and adjacent fjord systems, reaching its greatest density
at 250 m; however, specimens are rare above 100 m, and
only very infrequently found at the surface (Mackie and
Mills. 1983; Mackie, 1985). They are too fragile to be
collected and brought to the surface in nets. Thus, op-
portunities to study them have been few. In 1984 we ob-
tained enough specimens for a study of their prehensile
tentilla (Mackie et at.. 1988). In 1990 and 1991 we ob-
tained five more specimens, on which this account is
based. Having so few specimens necessarily limited the
scope of the study, but they were in good physiological
condition, and there is no reason to doubt the generality
of the findings.
Materials and Methods
Specimens of Euplokamis dunlapae Mills 1987 were
collected off the dock at the Friday Harbor Laboratories
and kept in wide-mouthed glass bottles at 8°C until used.
Material for electron microscopy was fixed in 2.5% glu-
taraldehyde in 0.4 M Millonig's phosphate buffer at pH
7.4 for 1 h at room temperature, rinsed, and osmicated
248
GIANT AXONS AND ESCAPE SWIMMING
249
Figure 1. Eitplokamix Jitnlapae swimming in an aquarium, oral end
up. Scale bar: 5 mm.
in 1% osmium tetroxide in the same buffer at 4°C for 1
h. Specimens invariably disintegrated during fixation de-
spite every precaution. Fortunately, intact fragments of
comb rows, along with some attached underlying tissue,
could be retrieved from the debris and processed for elec-
tron microscopy. The tissue was dehydrated through
graded ethanols and propylene oxide and embedded in
Epon 8 1 2. Thick sections were stained with Richardson's
stain. Thin sections were treated with uranyl acetate and
lead citrate. Because the tissues were extremely fragile,
we could not prepare whole mounts of living material for
examination by Nomarski or phase contrast microscopy.
Figure 2, showing giant axons in an intact, living animal,
was taken through a dissecting microscope illuminated
with a double substage mirror, with the mirror angles ar-
ranged to give shadows along the edges of the axons.
Behavior of free-swimming specimens was observed in
a 15-1 tank, illuminated from the sides, and with a dark
background. A Sony CCD video camera (HVM-200), fit-
ted with a Nikon 105 mm macro lens, was used in con-
junction with a VCR with frame-by-frame playback for
analysis of responses. Recordings of electrical activity were
made from specimens pinned down in a Sylgard-lined
Pyrex pie dish placed on top of a doughnut-shaped Cam-
bion cooling stage, which allowed light to enter from be-
low. Temperature was maintained at 12°C. Fine polyeth-
ylene suction electrodes were attached directly to the body
surface using minimal suction to keep them attached.
Signals were amplified with capacity-coupled amplifiers
and displayed on a digital oscilloscope; conventional ex-
tracellular recording procedures were used. Stimuli were
delivered through paired metal wires insulated to near the
tip. Movement of comb plates was monitored; we used a
ruby laser (Spectra Physics Model 1 55) to project a narrow
beam of light across the comb row, and a photomultiplier
(International Light 270C) to detect purturbations of the
beam caused by the ciliary movement. Laser monitoring
of ciliary beating in molluscan veligers is described by
Arkett el al. (1987). The method, as used here, allowed
us to distinguish forward from reverse power strokes by
the polarity of the waves recorded.
Histology and Ultrastructure
The comb rows of Euplokamis resemble those of other
ctenophores as reviewed by Tamm (1982). Each plate
consists of thousands of cilia springing from "polster"
cells, which are packed with large mitochondria. We have
confirmed that gap junctions are present between the
polster cells as first reported by Hernandez-Nicaise ( 1 974).
The comb plates are richly innervated by fine nerve fibers
running among the bases of the polster cells. These fine
neurites have diameters in the range of 0.7-2.5 /urn (X
= 1.2, SD = 0.4, n = 13). They appear to represent part
of the general ectodermal nerve plexus. This system is
well known in Pleurobrachia from the investigations of
Hernandez-Nicaise (1973a, 1974). Moss (1986) has shown
that ciliary reversals in Pleurobrachia are mediated by a
diffuse conduction system running in the ectoderm, pre-
sumably the nerve plexus described by Hernandez-Nicaise
in whole or part.
Where Euplokamis differs from other known cteno-
phores is in its possession of bipolar giant axons running
along each comb row. These can be seen under the dis-
secting microscope in the intact, living animal (Fig. 2),
and at higher magnifications in thick epon sections (Fig.
3) and electron micrographs (Fig. 4). They are present in
all eight comb rows and run roughly down the midline
of each comb row from one end to the other. Their cell
bodies lie between the comb plates, and there appears to
be one cell body between each pair of comb plates. They
evidently form a chain of cells arranged end to end, with
250
G. O. MACKIE ET AL.
Figure 2. Surface view of a sub-tentacular comb row in living animal, showing four comb plates whose
cilia are laid down flat on the surface, and giant axon chain (g). Scale bar: 100 jim.
Figure 3. One (1.0) fim section cut transversely but at a slight angle through a comb plate showing
massed cilia (dark mass, upper right) arising from polster cells, and giant axons (g). Gametes are seen in the
underlying endodermal canal. Scale bar: 50 /jm.
some overlap. In typical sections, one or two axon profiles,
but not more, are seen, showing that the neurites must
be quite short, probably less than 400 ^m. In their thicker
regions, they show diameters of approximately 8.5-12.0
Mm (X = 9.7, SD = 1.02, n = 10), but they taper towards
the ends. Allowing for shrinkage during fixation and
embedding, the axons are probably at least 12 /urn in the
living animal, which is also suggested by measurements
GIANT AXONS AND ESCAPE SWIMMING
251
Figure -4. Low power electron micrograph showing cross sections of
giant axons (g) surrounded by polster cell bases. Scale bar: 2 ^m.
on living preparations. There appears to be only one nu-
cleus per cell; it is long and thin and causes only a slight
swelling in the cell body region.
The axoplasm of the giant axons is remarkable for its
richly developed smooth endoplasmic reticulum. which
appears to form a continuous network of fine canals
throughout the entire cell. The axoplasm also contains
conspicuous bundles of microtubules. Rows of 100-nm
dense-cored vesicles have been seen associated with these
bundles (Fig. 5). The area around the nucleus does not
differ markedly from other parts of the axoplasm. There
is little indication of active protein synthesis. The nucleo-
lus is not prominent, and there is little rough endoplasmic
reticulum. Mitochondria and Golgi bodies (one is seen
in Fig. 6) are distributed along the whole length of the
axon.
We have seen no gap junctions in the nervous system
(nor are there any reports of such in the ctenophore lit-
erature), but we have observed synapses between neurites
of all sizes. There are synapses between fine neurites and
giants, and between giants in areas of overlap (Fig. 6).
Synapses also occur between these neurons and polster
cells (Fig. 7). The synapses resemble those seen in the
tentilla of Euplokamis (Mackie el al.. 1988), as well as
those described in other ctenophores. They are charac-
terized by "presynaptic triads" (Hernandez-Nicaise,
1973b, 1974) consisting of a mitochondrion embraced by
a flattened ER cisterna with an accompanying row of small
(50 nm) synaptic vesicles.
Behavior of Free-Swimming Specimens
When left to its own devices in a large tank, Euplokamis
shows bouts of forward "cruising" swimming, interspersed
with periods of quiescence. During cruising, ciliary me-
tachronal waves run down the comb rows with a frequency
of about 5 Hz, driving the animal forward (mouth lead-
ing). A specimen 15 mm long cruises at about 2 cm • s '.
The tentacles trail behind, partially extended and with
their tentilla coiled. When swimming stops, the body
swings around passively so that the mouth points up and
the tentacles hang down. The animal can hang in this
posture with its cilia either arrested or beating irregularly
and infrequently. Swimming animals will go into reverse
if they hit the walls of the tank, but otherwise they swim
forward steadily in the cruising mode.
Stimulation by touch or by an electric shock on any
part of the surface alters the pattern of swimming. If the
stimulus is applied at the front, for instance on the lips,
the animal rapidly changes the direction of the ciliary
power stroke in all eight comb rows, goes into reverse,
and swims backward for one or two body lengths, pauses,
and then resumes forward cruising behavior. During
backward swimming, velocities of about 4 cm-s~' were
observed, with elevated metachronal beat frequencies.
Using frame-by-frame playback on the VCR, the first signs
of interruption of the ciliary beat pattern were generally
seen within four frames (< 1 30 ms) following the stimulus,
with the actual change in the direction of movement oc-
curring one or two frames (<67ms) later.
When stimuli are applied at the rear (statocyst) end,
the animal responds by accelerated forward swimming:
within four frames (<130 ms) the metachronal rhythm
suddenly increases as in the backward escape response,
but with no change in the direction of the power stroke,
so the animal darts forward rapidly at a velocity of about
5.5 cm • s~'. After several seconds, it slows to the normal
cruising speed.
During both backward and forward escape swimming,
the tentacles contract. The first signs of contraction are
seen at about the same time as the first signs of change in
the pattern of ciliary beating.
Animals that have been left to swim around the tank
without interference respond to stimuli with great alacrity,
but after repeated stimulation, responses become less in-
tense and latencies tend to increase.
Electrophysiology
Propagation along the comb rows
If a recording electrode is placed directly over the giant
axon chain at any point along its length, and an electrical
stimulus is delivered further along the comb row, a pattern
252
G. O. MACK.IE ET AL
dcv
mt
•
.
i • , -
j&j
%
v
*
r - * a.. w<
-ser
•fie
1
- ' r . „ i/(
7
-% l-fSj
Figure 5. Axoplasm ol giant axon cut longitudinally showing dense-cored vesicles (dcv) aligned along
microtubule bundle (mt) and richly developed smooth endoplasmic reticulum (ser). Scale bar: 0.5 ^m.
Figures 6, 7. Synapses (arrowheads) between giant axons (6) and from giant axon to polster cell (7).
Scale bars: 0.5 um.
of electrical potentials is recorded, which has two readily
separable components (Fig. 8). The first component is a
small ( 150 ^V), sharp, spikey event conducted at 51-56
cm -s"1 (X = 53.4, SD = 1.95, n = 10). This event is
presumably the extracellular correlate of an action poten-
tial propagated in the giant axon chain. The signal is lost
when the recording electrode is moved even slightly to
one side or the other of the giant axon. The second com-
ponent is a larger, complex, and variable series of positive
and negative-going potentials that presumably represent
responses in the ciliated polster cells. This component
may be brief or it may take the form of flurries lasting
several hundred milliseconds. We were not able to analyze
these events in detail, but assume that they include
GIANT AXONS AND ESCAPE SWIMMING
253
8
Figure 8. Recording with an electrode placed directly over the giant
axon row. two successive sweeps in same position. Following electrical
stimulation further along the comb row in the aboral direction (black
spot marks shock artifact), a rapidly conducted, sharp, spikey event (ar-
rowhead) is recorded, followed by a much larger, complex, and variable
series of potential changes. Scale bar: 10 ms, 200 j/V.
summed synaptic events and regenerative events in the
polster cells, which, with some of their cilia, were partially
engulfed by the electrode. The large biphasic events, which
reach amplitudes in the millivolt range, would presumably
correspond to the calcium spikes described by Moss and
Tamm (1986, 1987) in Pleurobrachia.
The interval between the arrival of the giant nerve spike
at the recording electrode and the start of the polster cell
depolarization is quite variable, ranging from 5 to 20 ms
(X = 12.2 ms, SD = 5.5, n = 7). This may mean that the
polster cell depolarization is not initiated at the recording
site, but is propagated from an unknown and more or
less distant site. Consequently, unless the nerve spike is
actually recorded, it is impossible to give an accurate or
even consistent estimate of conduction velocity in the
nervous system. In spite of this, measurements of the la-
tency between stimulus and polster cell response can give
a rough figure (here termed "apparent conduction veloc-
ity") serving for comparative purposes in any one location.
These estimates assume that conduction always takes
place along the most direct route measured between stim-
ulus and recording points, but this too needs verification.
Apparent conduction velocities along the midline of
the comb row were reduced to 75% of their original value
after a single cut through the giant axon chain. Apparent
conduction velocity decreased progressively with further
cuts at different levels, reaching 55% of the original value
after the axon chain had been destroyed over most of its
length. At this point, it must be assumed that excitation
was travelling in the general epidermal nerve plexus. The
apparent conduction velocity along the comb rows after
destruction of giant axons is close to the value obtained
for conduction across the ectoderm in regions devoid of
giant axons, the mean velocity being 12 cm-s~' (SD
= 3.07, n = 1 7, three specimens). Incisions into the comb
row tissue did not reduce apparent conduction velocities
unless they intersected the giant axons. The fact that sev-
eral incisions must be made through the giant axon chain
at different levels along the comb row to bring the apparent
conduction velocity down to the non-giant velocity is
consistent with the idea that impulses can enter the giant
axons from a diffuse nerve net at numerous points along
the giant axon chain.
Simultaneous recording of polster cell depolarizations
and ciliary beat patterns
Animals pinned down in a small dish continue to show
normal cruising swimming with a metachronal wave fre-
quency of about 5 Hz (Fig. 9). Using the laser beam
method of monitoring ciliary beating, we could, in a few
cases, detect the exact moment at which this pattern was
interrupted following stimulation and determine the in-
terval between this change and the onset of the electrical
response in the polster cells (Figs. 10, 11). The giant axon
spike is not seen in these records, as the recording electrode
was not placed directly over the midline of the comb row.
In Figure 10, following stimulation at the oral end of a
comb row, the cilia switched to the reverse pattern of
beating 122 ms after the stimulus, and 85 ms after the
start of the electrical response recorded from polster cells
adjacent to the laser-monitoring site. Beat frequency was
25 Hz immediately after the change of gait, but the fre-
quency declined quickly, and the cilia shortly switched
back to beating in the forward cruising mode at 5 Hz.
The response latency of 122 ms recorded by this method
is consistent with the value of < 130 ms obtained from
the video-playback analysis.
When the stimulating and recording positions were re-
versed and a shock was delivered at the aboral end of the
comb row, a forward escape swimming response was ob-
tained (Fig. 1 1 ). Regular, slow metachronal beating
changed suddenly to rapid forward beating starting 143
ms after the shock, and 107 ms after the first component
of the electrical response recorded from adjacent polster
cells. After two beats at the equivalent of 17 Hz, the fre-
quency declined to 8 Hz and later (off the record) to 5
Hz, as normal cruising was resumed. We could not repeat
these experiments enough times to allow a proper statis-
tical analysis and cannot say, therefore, whether the re-
sponse latency is consistently shorter in the case of reverse
escape swimming than in forward, nor how this may vary
254
G. O. MACKJE ET AL
J
10
11 '•
y
Figures 9, 10, and 11. In each figure, the upper trace is a record of
electrical activity from a comb plate while the lower is a laser beam
record of the ciliary beating at the same spot. Figure 9 is a control (no
stimulus). Figure 10 shows the response to a shock delivered orally of
the recording electrode. Small upward events on the laser record corre-
spond to reverse ciliary beats. Figure 1 1 shows the response to shock
delivered aborally. Small downward events on the laser record correspond
to fast forward beats. Spots mark shock artifacts. Scale bars: 100 ms,
500 nV (upper trace) (9); SO ms, 500 ^V ( 10); 50 ms, 200 ^V ( 1 1 ).
with position along the comb row. In both cases, however,
the response latency can clearly be less than 1 50 ms.
The cilia generally appear to switch directly from one
mode of beating to another without a break, but in some
cases a short period of inactivity was observed before the
new pattern emerged. During these periods, the cilia ap-
peared to be in the "laydown" position described for
Pleurobrachia (Moss and Tamm, 1986), but this needs to
be verified.
Comparison with other species
We know of no previous reports of giant axons in the
comb rows of ctenophores. We have looked at living spec-
imens of Pleurobrachia bachei and Beroe sp. using the
optical arrangement that enabled us to see the giant axons
in Euplokamis (Fig. 2) and could see no comparable
structures. We have cut some sections of Pleurobrachia
and examined them under the EM with the same result,
confirming Hernandez-Nicaise ( 1973a, 1974), who found
only small-diameter neurites. Electrophysiological re-
cordings from Beroe, made in the same way and at the
same temperature ( 12°C) as Euplokamis. show the com-
plex responses associated with excitation of the polster
cells but no preceding neural event. Presumably the nerves
conducting the excitation are too small and scattered to
give a clear extracellular signal. The response latency is
also much longer than in Euplokamis, with apparent con-
duction velocities lying in the range of 21-25 cm-s~'.
This is similar to the mean value of 25 cm • s~' recorded
for Beroe ovala at 22°C by Hernandez-Nicaise et al.
(1980). Apparent conduction velocities in Pleurobrachia
are even slower, in the range of 1 1-16 cm • s~'. It would
appear that rapid conduction is peculiar to Euplokamis
and is presumably made possible by the giant axons found
in this species.
Discussion
Giant axons have evolved in many invertebrate groups
as mediators of rapid responses serving either for escape
or food capture. Such rapid movements are generally
muscular, but there is no reason why a streamlined animal
swimming by means of powerful cilia should not have
undergone selection for fast pathways mediating its lo-
comotory responses. Such appears to be the case with
Euplokamis. While many details of the ciliary control
mechanism remain to be elucidated, there can be little
doubt that the giant axons have evolved to bring about
rapid changes in ciliary beat frequency and direction in
the context of escape. The finding of giant axons in a
ctenophore, though interesting, is not likely to require
any drastic reconsideration of ctenophore phylogeny and
relationships. We agree with Bullock (1984) that giant
GIANT AXONS AND ESCAPE SWIMMING
255
fibers have probably evolved independently in many
groups.
The giant axons appear to be fairly short, thick, mono-
nucleate structures forming a chain with some overlap.
They synapse with each other, suggesting that they con-
stitute a single pathway. They also synapse with elements
of the diffuse nerve net, and can be regarded as a spe-
cialized pathway within this net, recalling a similar situ-
ation in certain hydrozoan coelenterates (Mackie, 1989).
Ultrastructurally, the axons are rather unusual. They have
an extremely rich smooth endoplasmic reticulum that ap-
pears to form a continuous canal network throughout the
entire cell. Such a system could serve for intracellular
transport (Droz et al., 1975), but we also see EM images
of vesicles lined up along microtubule bundles (e.g., Fig.
5), which is suggestive of mechanoenzyme-driven trans-
port (Vallee et al., 1989).
We are not yet in a position to explain exactly how
ctenophores control the frequency of ciliary beating or
the direction of the power stroke, though much progress
has been made in recent years with Pleurobrachia. In this
genus, the process can be explained without resorting to
a dual innervation model (Moss and Tamm, 1986). Mod-
erate depolarization of the polster cells is associated with
increased rates of beating in the normal (forward) direc-
tion. Larger depolarizations sufficient to cause a regen-
erative response (calcium spike) are associated with arrest
(laydown) and with accelerated beating in the reverse di-
rection. The comb plates can pass fairly abruptly from
accelerated reverse to accelerated forward beating, al-
though a period of intermediate beating may intervene.
Thus, depending simply on the number and time rela-
tionships of input events at a single set of synapses, the
polster cells could be maintained at any level of depolar-
ization or spike frequency required for the various loco-
motory gaits. Moss and Tamm, however, found evidence
of two types of excitatory post-synaptic potentials, sug-
gesting that there may actually be two functionally distinct
neuronal excitatory pathways, one associated with a short
latency response and moderate depolarization of the
polster cells, and the other with a more delayed response
and facilitating excitatory post-synaptic potentials that
generate spikes and cause ciliary arrest or reversal. Recall,
in this context, that three subsets of neurites were described
in the ectodermal plexus of Pleurobrachia from ultra-
structural and pharmacological studies (Hernandez-Ni-
caise, 1974). Pharmacological evidence of two subsets was
also presented by Satterlie (1978).
At first sight, our physiological findings on Euplokamis
favor the dual innervation theory because they show that
the direction in which the cilia beat depends on the di-
rection from which the excitation is coming. The easiest
way of explaining this observation would be to assume
that there are two separate diffuse nerve nets, one receiving
sensations from the front and one from the back, and
exciting the comb plates in different ways. If so, however,
there should be two sets of giant axons, one associated
with each net, but this is not the case. The giants appear
to constitute a single series in each comb row, adjacent
giants being interconnected by synapses. If this is the case,
then there is only one cilio-motor innervation, and the
giants are part of it, providing the final common pathway
for information from all parts of the animal. In this sce-
nario, the way the comb plates respond might depend on
impulse frequency differences associated with the exci-
tation of receptors at the two ends, transmitted through
the common excitatory pathway. Frequency coding in a
single conducting system provides the basis for two sorts
of behavior in the sea anemone Actinia (McFarlane and
Lawn, 1991). However, other explanations are possible,
and as the evidence needed to decide between them is not
yet available, further speculation on this question is in-
appropriate. Very probably, however, the control system
in Euplokamis will prove to be a modified version of the
Pleurobrachia system, and as Pleurobrachia is a rugged
and common surface-living species, it is probably better
suited for use in exploring these questions than our fragile
and elusive midwater species.
Acknowledgments
We thank Dennis Willows, Director of the Friday Har-
bor Laboratories, for space and laboratory facilities and
Colin Hermans for help collecting specimens. Tony Moss
kindly made available portions of his doctoral dissertation.
This research was funded by the Natural Sciences and
Engineering Research Council of Canada.
Literature Cited
Arkett, S. A., G. O. Mackie, and C. L. Singla. 1987. Neuronal control
of ciliary locomotion in a gastropod veliger (Calliostoma). Bio! Bull
173:513-526.
Bullock, T. H. 1984. Comparative neuroeothology of startle, rapid es-
cape and giant fiber-mediated responses. Pp. 1-13 in Neural Mech-
anisms of Startle Behavior, R. C. Eaton, ed. Plenum. New York.
Droz, B., A. Rambourg, and H. L. Koenig. 1975. The smooth endo-
plasmic reticulum: structure and role in the renewal of axonal mem-
brane and synaptic vesicles by fast axonal transport. Brain. Res 93:
1-13.
Hernandez-Nicaise, M.-L. 1973a. Le systeme nerveux des Ctenaires
I. Structure et ultrastructure des reseaux epitheliaux. Z Zellforsch.
137: 223-250.
Hernandez-Nicaise, M.-L. 1973b. The nervous system of ctenophores
III. Ultrastructure of synapses. J Neurocytol, 2: 249-263.
Hernandez-Nicaise, M.-L. 1974. Systeme nerveux et integration chez
les Ctenaires: etudes ultrastructurale et comportementale. Doctoral
dissertation. Universite Claude-Bernard (Lyon I).
Hernandez-Nicaise, M. L., G. O. Mackie, and R. W. Meech.
1980. Giant smooth muscle cells ofBeroe. J. Gen. Physiol. 75: 79-
105.
256
G. O. MACKJE ET AL.
Mackie, G. O. 1985. Midwater macroplankton of British Columbia
studied by submersible PISCES IV. / Plankt Res. 7: 753-777.
Mackie, G. O. 1989. Evolution of cnidarian giant axons. Pp. 395-407
in Evolution of the First Nervous Systems. P. A. V. Anderson ed.
Plenum, New York.
Mackie, G. O., and C. E. Mills. 1983. Use of the PISCES IV submersible
for zooplankton studies in coastal waters of British Columbia. Can.
J. Fish. Aijitat. Sa. 40: 763-776.
Mackie, G. O., C. E. Mills, and C. L. Singla. 1988. Structure and
function of the prehensile tentilla of Euplokumis (Ctenophora, Cy-
dippida). Zoomorphology 107: 319-337.
McFarlane, I. D.. and 1. D. Lawn. 1991. The senses of sea anemones:
responses of the SSI nerve net to chemical and mechanical stimuli.
Hydrobiologia 216/217: 599-604.
Mills, C. E. 1987. Revised classification of the genus Euplokumis Chun,
1880 (Ctenophora: Cydippida: Euplokamidae n. fam.) with a de-
scription of the new species Euplokamis diinlapae. Can. J. Zool 65:
2661-2668.
Moss, A. G. 1986. The control of ciliary comb plate beating in the
ctenophore Pleurobrachia. Doctoral dissertation. Boston University.
Moss, A. G., and S. L. Tamm. 1986. Electrophysiological control of
ciliary motor responses in the ctenophore Pleurobrachia. J Comp.
Physiol. 158:311-330.
Moss, A. G., and S. L. Tamm. 1987. A calcium regenerative potential
controlling ciliary reversal is propagated along the length of ctenophore
comb plates. Proc. Nail. Acad. Sci. USA 84: 6476-6480.
Satterlie, R. A. 1978. Feeding mechanisms in the ctenophore Pleu-
robrachia pileus. Bio/. Bull. 155:464.
Tamm, S. L. 1982. Ctenophora. Pp. 266-358 in Electrical Conduction
and Behaviour in 'Simple' Invertebrates. G. A. B. Shelton, ed. Clar-
endon Press, Oxford.
Tamm, S. L., and A. G. Moss. 1985. Unilateral ciliary reversal and
motor responses during prey capture by the ctenophore Pleurobrachia.
J Exp. Biol. 114:443-461.
Vallee, R. B., H. S. Shpetner, and B. M. Paschal. 1989. The role of
dynein in retrograde axonal transport. Trends Neurosci. 12: 66-70.
Reference: Biol. Bull. 182: 257-264. (April, 1992)
Phase Shift of a Tidal Rhythm by Light-Dark Cycles
in the Semi-Terrestrial Crab Sesarma pictum
MASAYUKI SAIGUSA
Okayama University. College of Liberal Arts & Sciences, Tsushima 2-1-1, Okayama 700, Japan
Abstract. The larval release activity of the semi-terres-
trial crab Sesarma pictum was monitored for three-week
periods under laboratory conditions of constant and cyclic
light. Under conditions of constant dim light, the rhythm
for the first ten days was unimodal (larval release just
after the nocturnal high tide) and then became bimodal
(no apparent synchrony with the tides or with other
members of the population) for the remainder of the ex-
perimental period. On the other hand, in photoperiods
similar to those in the field, the rhythm was maintained;
the phase was bimodal and the timing of larval release
was delayed 1 -2 h from the predicted times of high water
in the habitat. When the photoperiod was advanced or
delayed, the tidal rhythm was phase-shifted accordingly.
The photoperiod does entrain the release rhythm to bi-
modal tidal cycle. So the phase-shift of a tidal rhythm by
24-h LD cycles is a very difficult phenomenon to explain.
Introduction
In their natural habitats, intertidal and estuarine ani-
mals are exposed, not only to the day-night cycle (24 h),
but also to the rhythmic ebb and flow of the tides, which
include 12.4-h, 24.8-h, and 15-day components. Having
adapted to such an environment, marine organisms often
show a complex activity pattern with circadian, circatidal,
and circa-semilunar frequencies; the dominant rhythmic-
ity is circadian in some species, and circatidal or semilunar
in others (see reviews by Neumann, 1981; DeCoursey,
1983).
Compared with the terrestrial animals, information
about the biological timing systems is relatively limited
in marine animals. This paucity of information is partly
due to the complexity of environmental cycles. In addi-
tion, most biological timings have been investigated in
locomotor activities. The noisy nature of these activities.
Received 27 March 1990; accepted 21 January 1992.
and the individual variability in the responses to environ-
mental cycles, have made the analyses difficult in most
aquatic animals. For example, when the locomotor activ-
ity of the fiddler crab Uca crenulata was monitored in
constant light or 24-h light-dark conditions, half of the
experimental crabs only showed a rhythmic activity; the
activity of the remaining half was random (Honegger,
1973).
Clearly demarcated biological rhythms have been re-
ported in the swimming activity of some marine crusta-
ceans (Enright, 1963, 1972). The records of their activity
have demonstrated predominantly circatidal rhythms that
were not affected by light-dark cycles in the laboratory.
Animals do entrain or respond to simulated tidal stimuli,
such as wave action in the isopod Excirolana (Enright,
1965), or cycles of hydrostatic pressure in the amphipods
Synchelidium (Enright, 1962) and Corophium (Morgan,
1965).
Precise biological timings also often develop in repro-
ductive phenomena, and this has been observed in a va-
riety of marine animals, including the polychaete Platy-
nereis (Hauenschild, 1960), the intertidal midge Clunio
(Hashimoto, 1976; Neumann, 1976), and many species of
Crustacea (Branford, 1978; DeCoursey, 1979, 1983). The
larval release behavior of the estuarine terrestrial crab Se-
sarma haematocheir is also synchronized with environ-
mental light and tidal cycles, showing a unimodal tidal
rhythm that coincides with the times of nocturnal high
water (Saigusa, 1982, 1985). A phase jump is involved in
the timing process, so that this tidal rhythm appears at
1 5-day intervals. Experimental analyses indicated that the
timing is endogenously controlled, and that the phase of
the rhythm can be shifted by artificial 24-h light-dark (LD)
cycles (Saigusa, 1986). An important problem is the timing
mechanism underlying the tide-synchronized biological
rhythm, the phase of which is shifted by day-night cycles.
In this paper, ovigerous females of Sesarma pictum were
used for experiments, because larval release activity in
257
258
M. SAIGUSA
B
controller
event
(with switch)
recorder
original record (15 July. 1990)
Time of day (h)
£3 =
I
-= '
*rL
4
-u—
• !
'
b
a —
t .
b
TT
controller
Figure 1. System for recording the larval release activity of Sesarma pictum females. A. The apparatus
used to detect the time of day of larval release, w: fine wire, pc: plastic cage to confine an ovigerous crab, s
and r: infrared source and receiver (E3S-2E4, Omron Co. Ltd., Japan). The glass beaker illustrated here is
the larger one (13.5 cm diameter; see text). B. An example of an original record showing 4 out of 17 females
releasing their larvae (a part of the data illustrated in Figure 2). The output of the sensor unit was monitored
by an event recorder (R17-H12T, Fuji Electric Co. Ltd., Japan) through a controller (S3S-A-10, Omron).
a: time of larval release by a female, b: time when that female was removed from the experimental chamber
with her beaker. Simultaneously, a new beaker with a plastic cage that confined another ovigerous female
was placed in each apparatus.
this species has a tidal rhythm with a bimodal phase (12.5-
h period) that has no apparent circadian component. The
present paper asks whether this tidal rhythm is also af-
fected by the day-night cycle, and it discusses possible
mechanisms for this influence.
Materials and Methods
Sesarma pictum inhabits banks above the water's edge
in the intertidal zone. Male and female crabs spend the
winter hibernating in burrows dug into the bank, but they
become active in the latter half of April. In early summer
(June-July), the females incubate their eggs in their folded
abdomens, where the embryos are ventilated by move-
ments of the pleopods. When embryonic development is
completed, the females enter the water to liberate their
zoea larvae. The larval release behavior has not been ob-
served in the field because it is not carried out at the water's
edge, as is that of S. haematocheir (Saigusa, 1982, 1985).
Rather, the ovigerous females of 5. pictum enter the water
at about high tide, and then soon disappear in the depths;
possibly they release their larvae in the water near the
shore.
Ovigerous females were collected from the seacoast at
Kasaoka, Okayama Prefecture. The crabs occur in narrow
crevices between rocks on the bank; they were stimulated
to emerge with a thin stick, and were then captured by
hand. Once suitable numbers had been collected, they
were brought into the laboratory and placed in aquaria
(70 cm long, 40 cm wide, and 25 cm high). The aquaria
had a shallow pool of diluted seawater (salinity at about
20%o) at the bottom, and hiding places (moistened with
fresh water) made of boards set above the surface of water.
The crabs were fed every few days. The experimental
rooms were equipped with controlled light and temper-
ature. Temperature was at 23 ± 1.5°C; luminous intensity
was about 700-1200 lux at the floor with the light on,
and less than 0.01 lux in the dark phase. Only two ex-
periments were carried out under continuous light, and
in these cases the luminous intensity was at 0.5-1.0 lux,
or 100-300 lux, respectively.
The eggs (i.e.. embryos with egg capsules) of each
female were checked by eye every day, and those crabs
carrying embryos that seemed likely to hatch within a
few days were individually set in a recording apparatus
placed in the same room (Fig. 1A). With this apparatus,
the time of larval release could be monitored without
any change in the ambient lighting conditions. The day
of hatching is difficult to predict; the only indication
that hatching is imminent is the brownish green color
of the embryos (mainly caused by yolk consumption).
The larval release behavior of S. pictum was observed
in the laboratory, and was generally not as vigorous as
that of S. haematocheir. The female repeatedly flexed
her abdomen inward and made associated movements
of the pleopods bearing embryos. These pumping
A TIDAL RHYTHM OF THE SEMI-TERRESTRIAL CRAB
259
1990
3 15
12
18
Time of day
6
12
18
25
HW2
SS
HWi SR
HW2
SS
SR
Figure 2. Time of day of larval release monitored under a regime of continuous light (LL: 0.5-1.0 lux)
and no tidal influence in the laboratory. All of the ovigerous females were collected in the field on 3 July
1990; black dots indicate the time of day of larval release by those females. For comparison, environmental
cycles in the field are characterized by the times of sunset (SS) and sunrise (SR), by the curves connecting
predicted time of day of high tides (HW, and HW2), and by the phase of the moon (O: full moon, 9: the
last quarter of the moon, •: new moon). The entire record is duplicated on the right and displaced upwards
one day, so that each day's data can be matched with those of the following day. Eighty (80) females were
used in this experiment.
movements swept clouds of newly hatched zoeae away
from the female.
The larval release recording system consisted of a sensor
unit (infrared source and receiver) placed inside the ex-
perimental room, and a controller unit placed outside.
Each ovigerous female was confined in a separate, small
plastic cage (6 cm in diameter and 8 cm in height) with
many holes drilled in the bottom and sides. As Figure 1 A
shows, the cage was suspended by a fine wire from the
rim of a glass beaker containing diluted, clean seawater
(salinity at about 20%o). The experimental procedures for
monitoring the larval release activity of S. picturn were
basically the same as those for 5. haematocheir (Fig. IB),
and they have already been described elsewhere (Saigusa,
1986).
For each female, one of two kinds of glass beakers was
selected to hold her plastic cage: a bigger one ( 1 3.5 cm in
diameter), or a smaller one ( 1 1 cm in diameter). The se-
lection was made to ensure that the photoelectric switch
would respond and was based on the size of the egg sponge
carried by the female. Moreover, because the ovigerous
female of S. pictum has a smaller carapace (1.6-2.5 cm)
than S. haematocheir, the amount of seawater in the larger
beaker was reduced to 0.5 1, and to 0.3-0.4 1 in the smaller
one. The number of ovigerous females used in each ex-
periment is described in the figure legends or in the text.
(Some females were released into the aquaria before being
confined to cages. Because the time of the release of those
females was not monitored, such releases were not in-
cluded in the figures.) The animals were not fed after they
had been confined in the recording apparatus.
Results
The females used in the present experiments were ran-
domly collected from the field; thus, some crabs seemed
ready to release larvae within a few days, whereas others
carried eggs that seemed to have commenced incubation
just one or two days before. The larval release by these
females was completed within three weeks after the col-
lection, suggesting that females incubate their embryos
for about 2-3 weeks.
260
M. SAIGUSA
12
Time of day
18 24
12
18
19881
Figure 3. Daily timing of larval release by Sesarma pictum monitored under the conditions of a 24-h
light-dark (LD) cycle in the laboratory. Forty crabs were used. Date of collection: 16 June 1988. Vertical
lines indicate the times of light-off and light-on in the chamber, respectively. SS and SR are the times of
sunset and sunrise in nature. Diagonal curves (HW, and HW, ) connect the times of high water in the field.
Diagonal lines (RL, and RL2) are least squares regression lines fitted to each phase of the tidal rhythm. The
period length of each phase (r) is estimated from the slope (a) of its regression line (r = 24h + a). The slopes
of RL, and RL2 are 0.75 and 0.84, respectively. The regression lines are based on those data obtained after
the phase shift was considered to have been completed.
The first experiment was designed to determine whether
an endogenous component is involved in the larval release
activity and, if so, whether this component corresponds
to the day-night cycle or to tidal cycle at the local habitat.
For these purposes, the larval release activity of the pop-
ulation (80 specimens) was monitored under constant,
very dim light (LL) conditions for more than three weeks
following collection. As indicated in Figure 2, the larval
release activity persisted under these conditions. For the
first 10 days, the phase of the rhythm was unimdal (24.5-
h period) and the larval release roughly coincided with
the nocturnal high tides. The release rhythm then became
bimodal (12.5-h period), but no apparent synchrony with
the tides, or with other members of the population, ap-
peared for the remainder of the experimental period. A
similar experiment, with 40 individuals under stronger
luminous intensity ( 100-300 lux), also had the same ten-
dency (not illustrated).
To examine the effect of light regime, experiments were
then conducted with a 1 5-h light: 9-h dark photoperiod;
i.e., the phase of the cycle was set to be similar to that in
the field. The effects of cyclic light were markedly different
from those illustrated in Figure 2: the timing of the larval
release was closely correlated with the tidal cycles (high
water) for at least three weeks (Fig. 3). The phase of this
rhythm is clearly bimodal (12.5-h period). Least squares
regression lines fitted to each phase of the rhythm (RL,
and RL2 ) showed that the timing of release monitored in
the laboratory was delayed 1 -2 h from the predicted times
of high water in the habitat. The correlation of larval re-
lease with the tidal cycle in the habitat of the crabs con-
tinued throughout the experiment. Another experiment,
with 50 females and the same light conditions, was per-
formed on 1-20 July 1987; the results were the same as
those of Figure 3 (not illustrated). A comparison of Figures
2 and 3 suggests that a 24-h natural light-dark cycle main-
tains the normal tidal phase of a population rhythm.
Next, I examined whether the phase of the tidal rhythm
can be affected if the phase of the experimental cyclic
light is shifted from the natural day-night cycles. The fol-
lowing experiments were made to confirm this point
quantitatively. In Figure 4, the phase of the artificial day-
night cycle was advanced relative to the natural light cycle,
by 6.25 h at lights-on, and by 7 h at lights-off. After some
days had elapsed, the time of release shifted ahead of the
predicted high water curves. This suggests an advanced
A TIDAL RHYTHM OF THE SEMI-TERRESTRIAL CRAB
261
12
18
Time of day
24
12
18
1988 c
-6.25h-
Figure 4. Time of day of larval release monitored under a 24-h light-dark regime (LD 15:9), the phase
of which was changed by 6-7 h with respect to the natural conditions. Times of light-off and light-on in the
artificial light cycles are shown by vertical lines (light-on at 22:00, light-off at 13:00), and times of sunset
and sunrise are marked by the broken lines. RL, and RL: indicate least squares regression lines applied to
the new phase after the shift; i.e., the data from 29 June-13 July for RL!, and those of 30 June-17 July for
RL2. Other symbols are the same as in Figure 2. Collection of crabs: 23 June 1988. The slopes of RL, and
RL2 are 0.76 and 0.80, respectively. Fifty animals were used in the experiment.
phase-shift. The magnitude of the shifts in the two phases
were somewhat different; i.e., whereas the time difference
between HW, and RL, was 5-6 h, that between HW2 and
RL2 was 4-5 h. The phase differences between the rhythms
in both Figures 2 and 3, corresponded to the time lag
between the natural day-night and the artificial 24-h LD
cycles.
Another experiment was also meant to verify that the
magnitude of the phase-shift of the tidal rhythm is de-
pendent on the phase difference between natural and ar-
tificial day-night cycles. In this case, the experiment asked
whether the phase can be delayed. One hundred females
were used in these experiments, and the light regime was
shifted by 5 h at lights-on and by 5.7 h at lights-off. In
this experiment, the data suggest that, after about 1 0 days,
a delayed phase-shift occurred (Fig. 5). Another experi-
ment, in which 35 animals received the same treatment
(23 July to 7 August 1987), clearly demonstrated a similar
phase delay (not illustrated). Figure 5 also shows that the
phase of the population rhythm remained stable with re-
spect to the phase of high water, for at least the next 2-3
weeks, with no notable desynchronization of the individ-
uals. The time lag between HW, and RL, was 5.5-7 h,
and between HW2 and RL2 it was 5.5-6.5 h. The duration
of the phase delay could not be determined in these ex-
periments, because the number of the females incubating
the next clutch diminished.
The experiments described above were performed with
crabs collected on different dates. Uncertainties remained,
theiefore, about whether a light cycle can actually phase-
shift a tidal rhythm, and if so, whether advancing or de-
laying the photoperiod truly corresponds to the change
in the tidal rhythm. To meet this question, about 200
ovigerous females were collected from the field on 3-4
July 199 1 , and randomly separated in the laboratory into
two groups of similar size. One group was exposed to an
artificial 24-h LD cycle, the phase of which was similar
to the natural LD cycle (Fig. 6A); the other group was
exposed to an artificial LD cycle that was advanced 4-5
h from the natural light cycle (Fig. 6B).
In the control experiment (Fig. 6A), the larval release
occurred just after the time of high tides in the field. The
262
12
18
M. SAIGUSA
Time of day
24
12
18
5.7 h
Figure 5. Time of day of larval release monitored under a 24-h light-dark regime (LD 15:9), the phase
of which was delayed by 5-6 h from the natural light cycle. Times of light-on and light-off were 1 :00 and
10:00, respectively. All of the crabs were collected on 21 June 1989. Some of the females incubated a second
clutch in the laboratory, and the larval release activity of those crabs was also monitored. The estimated
slope of RL, and RL2 is 0.85 and 0.86, respectively. 3: the last quarter of the moon, •: new moon, C: the
first quarter. About 100 females were used in the experiment.
phase of this rhythm was clearly bimodal. These features
were the same as those in Figure 3. The crabs exposed to
the light cycle that had been advanced (Fig. 6B) also
showed a bimodal tidal rhythm, but a week had elapsed,
and the time of release shifted ahead of the time of high
tides. The time lag between the activity after 10 July and
high tide (HW2) showed an advancing phase-shift of
about 4 h.
No clear indication of a semilunar component (i.e., a
semi-monthly fluctuation in the number of females re-
leasing larvae per day) was found in these results. Neither
was a 24-h solar day (i.e., circadian) component detected,
at least in the activity pattern itself.
Discussion
The purpose of devising experiments in constant light
(e.g., Fig. 2) is to demonstrate a free-running rhythm.
Certainly the larval release was roughly correlated with
the time of nocturnal high tides for the first 10 days, and
then the rhythm became bimodal. However, no apparent
synchrony with the tides, or other individuals in the pop-
ulation, was seen for the latter half of the experimental
period; so no free-running rhythm was clearly evident in
Figure 2.
In most studies of rhythmic behavior, activities that are
carried out repeatedly by each individual are monitored
throughout the investigation. In this work, however, each
crab released larvae just once during a three-week exper-
imental period; so no free-running rhythm was evident.
A possible explanation of the data in Figure 2 is, therefore,
that the constant light increased the variability of the free-
running period in each individual, desynchronizing the
population rhythm.
The first question arising here is related to the envi-
ronmental cues that entrain the tidal rhythm of this spe-
cies. Circa-tidal rhythms are known to respond to stimuli
correlated with on-shore tides, and not to day-night cycles.
Enright (1965) showed that cycles of water turbulence
can effectively entrain the circa-tidal rhythm of the isopod
A TIDAL RHYTHM OF THE SEMI-TERRESTRIAL CRAB
263
Time of day
12
1991
Jul 3
9 5
Figure 6A. Time of day of larval release monitored under a 24-h
LD cycle (LD 15:9). the phase of which was similar to that of the field.
Date of collection: 3-4 July 1991. About 100 animals were used in the
experiment. Symbols were the same as in Figure 2.
Excirolana. Cyclical or non-cyclical changes of hydrostatic
pressure have been shown to cause behavioral responses
in the amphipods (Enright, 1962; Morgan, 1965). How-
ever, all the experimental data obtained in this study (Figs.
2-5 and 6A, B) have demonstrated that a light regime
actually takes part in the phase shift. The 24-h LD cycle
may be the zeitgeber of the Sesanna pictum tidal rhythm.
In the field, however, this rhythm is not likely to be en-
trained solely by the 24-h LD cycle.
The tidal rhythm of Sesanna haematocheir was en-
trained by 24.5-h artificial moonlight cycles administered
in the dark period of a 24-h LD cycle (Saigusa, 1988,
1989). In view of these studies, the S. pictum larval release
rhythm could be entrained by more than one environ-
mental cue. A 24-h LD cycle is one zeitgeber, but others
remain unknown. The habitat of 5. pictum is restricted
to the bank along the shoreline, so tidally correlated fac-
tors, such as the periodic fluctuations of water turbulence
on the shore, should also be considered.
The second, and central, question posed in this paper
is the effect of light on the tidal rhythm of 5. pictum. If
the results of Figures 4 and 5. and those of Figure 2, were
regarded as arising from substantially similar mechanisms,
there would be no need to assume a phase shift caused
by the environmental light cycle. In this case, one possible
explanation is that light cycles exert some superficial in-
fluence on the phase of the tidal rhythm irrespective of
timing mechanisms, causing abnormal phasing of the
rhythm. However, the results of Figure 6A and B would
completely deny such a possibility; a photoperiod does
entrain larval release rhythm to a bimodal tidal cycle.
Because a 24-h LD cycle can evoke a phase shift of the
tidal rhythm, then the difficulty is understanding the tim-
ing mechanism involved.
Many investigators have found that behavioral and
physiological events in marine organisms not only coin-
cide with the tidal cycle, but are also correlated with the
day-night cycle to produce activity patterns with simul-
taneous daily and tidal components (Naylor, 1958; Barn-
well, 1966; Palmer and Round, 1967; Honegger, 1973;
Benson and Lewis, 1976). Accordingly, recent interpre-
tations have included two types of internal clocks: one of
circatidal frequency, and the other of circadian frequency
(Naylor, 1958;Barnwell, 1966, 1968; Palmer and Round,
1967; Benson and Lewis, 1976; Webb, 1976). For ex-
ample, two internal clocks were proposed to explain noc-
Time of day
12
24
1991
Jul 3-
9 5
10-
20-
25
30-
HWl
-SR
SS
52h
4.1 h
Figure 6B. Time of day of larval release recorded in a 24-h LD cycle
(LD 15:9), the phase of which was advanced by 4-5 h from that of the
field. Date of collection was the same as in Figure 6A. Times of light-
on and light-off were at 0:00 and 15:00, respectively. One hundred and
five animals were used in the experiment.
264
M. SAIGUSA
turnal locomotion in the amphipod Talorchestia (Benson
and Lewis, 1976). One of these, a circadian clock, controls
the nocturnal phase of the activity, and the other, a circa-
tidal clock, inhibits the activity around the time of noc-
turnal high water.
The phenomenon reported in this paper, i.e., phase
shift of the tidal rhythm appearing under cyclic light (Figs.
4, 5, and 6B), could not be explained in terms of an in-
teraction between circa-tidal and hidden circadian
rhythms operating simultaneously within individuals. The
reason is based on the understanding that a circa-tidal
rhythm is only the expression of a circadian rhythm which,
as a result of an adaptation to marine environment, has
slightly modified its internal period and is responsible to
tide-correlated zeitgebers, too. Yet clearly, the circa-tidal
rhythm of S. pictum cannot be explained in terms of a
hypothesis requiring that those tide-correlated zeitgebers
affect a circadian (bimodal) rhythm directly, changing its
internal period to that of a bimodal tidal cycle.
The nocturnal release rhythm of S. haematocheir was,
therefore, accounted for by a hypothesis similar to the
mechanisms proposed by Pittendrigh and coworkers (Pit-
tendrigh, 1960, 1981; Pittendrigh and Bruce, 1959) (see
Saigusa, 1986 and 1988, for details). In that model, when
the driven oscillator is delayed until dawn, it leaps back
to dusk. However, if one wants to explain the effect of
light by the mechanism that was applied to S. haemato-
cheir, then one would have to assume that the driving
oscillator can control the phase of the driven oscillator
which has a different period. Perhaps such an assumption
is not realistic. Thus, the phase-shift of a tidal rhythm by
24-h LD cycle is a very difficult phenomenon to explain.
Acknowledgments
I thank Mr. A. Bettchaku, Faculty of Education, Oka-
yama University, who helped me to collect crabs in the
field.
Literature Cited
Barnwell, F. H. 1966. Daily and tidal patterns of activity in individual
fiddler crab (Genus Uca) from the Woods Hole region. Biol. Bull
130: 1-17.
Barnwell, F. H. 1968. The role of rhythmic systems in the adaptation
of fiddler crabs to the intertidal zone. Am. Zoo/. 8: 569-583.
Benson, J. A., and R. D. Lewis. 1976. An analysis of the activity rhythm
of the sand beach amphipod, Talorchestia quovana. J. Comp. Phvsiol.
105: 339-352.
Branford, J. R. 1978. The influence of daylength, temperature and
season on the hatching rhythm of Homarus gammams. J Mar. Biol.
Assoc. U.K. 58: 639-658.
DeCoursey, P. J. 1979. Egg-hatching rhythms in three species of fiddler
crabs. Pp. 399-406 in C\~clic Phenomena in Marine Plants and An-
imals. E. Naylor and R. G. Hartnoll, ed. Pergamon Press, Oxford.
DeCoursey, P. J. 1983. Biological timing. Pp. 107-162 in The Biology
of Crustacea. \'oi 7: Behavior and Ecology. F. J. Vernberg and
W. B. Vernberg, ed. Academic Press, Washington.
Enright, J. T. 1962. Pressure sensitivity of an amphipod. Science 133:
758-760.
Enright, J. T. 1963. The tidal rhythm of activity of a sand-beach am-
phipod. Z I'ergl. Physiol. 46: 276-313.
Enright, J. T. 1965. Entrainment of a tidal rhythm. Science 147: 864-
867.
Enright, J. T. 1972. A virtuoso isopod. Circa-lunar rhythms and their
tidal fine structure. J. Comp. Physiol. 77: 141-162.
Hashimoto, H. 1976. Non-biting midges of marine habitats (Diptera:
Chironomidae). Pp. 377-414 in Marine Insects. L. Cheng, ed. North-
Holland Publishing Company. Amsterdam.
Hauenschild, C. 1960. Lunar periodicity. Pp. 491-497 in Biological
Clocks. Cold Spring Harbor Symposia on Quantitative Biology, Vol.
25, New York.
Honegger, H.-W. 1973. Rhythmic motor activity responses of the Cal-
ifornia fiddler crab Uca cremilata to artificial light conditions. Mar.
Biol. 18: 19-31.
Morgan, E. 1965. The activity rhythm of the amphipod Corophium
volulalor (Pallas) and its possible relationship to changes in hydrostatic
pressure associated with the tides. / Anim. Ecol. 34: 731-746.
Naylor, E. 1958. Tidal and diurnal rhythms of locomotory activity in
Carcinus maenas (L.). J. Exp. Biol. 35: 602-610.
Neumann, D. 1976. Adaptations of chironomids to intertidal environ-
ments. Annu. Rev. Entomol. 21: 387-414.
Neumann, D. 1981. Tidal and lunar rhythms. Pp. 35 1-380 in Biological
Rhvthins. Handbook of Behavioral Neurobiology. Vol. 4, J. Aschoff,
ed. Plenum Press. London.
Palmer, J. D., and F. E. Round. 1967. Persistent, vertical-migration
rhythms in benthic microtlora. VI. The tidal and diurnal nature of
the rhythm in the diatom Hant:schia virgata. Biol Bull 132: 44-
55.
Pittendrigh, C. S. 1960. Circadian rhythms and the circadian organi-
zation of living systems. Pp. 159-184 in Biological Clocks. Cold Spring
Harbor Symposia on Quantitative Biology. Vol. 25, New York.
Pittendrigh, C. S. 1981. Circadian systems: general perspective. Pp.
57-80 in Biological Rhythms Handbook oj Behavioral Neurobiology.
Vol. 4, J. Aschoff, ed. Plenum Press, London.
Pittendrigh, C. S., and V. G. Bruce. 1959. Daily rhythms as coupled
oscillator systems and their relation to thermoperiodism and pho-
toperiodism. Pp. 475-505 in Photoperiodism and Related Phenomena
in Plants and Animals. R. B. Withrow. ed. American Association for
the Advancement of Science. Washington.
Saigusa, M. 1982. Larval release rhythm coinciding with solar day and
tidal cycles on the terrestrial crab Sesarma. Biol. Bull 162: 37 1 -386.
Saigusa, M. 1985. Tidal timing of larval release activity in non-tidal
environment. Jpn. J. Ecol. 35: 243-251.
Saigusa, M. 1986. The circa-tidal rhythm of larval release in the in-
cubating crab Sesarma. J Comp. Physiol. 159: 21-31.
Saigusa, M. 1988. Entrainment of tidal and semilunar rhythms by
artificial moonlight cycles. Biol. Bull. 174: 126-138.
Saigusa, M. 1989. A circadian rhythm in tidal environment: larval
release activity of the terrestrial crab Sesarma. Pp. 175-186 in Cir-
cadian Clocks and Ecology. T. Hiroshige and K. Honma, ed. Hok-
kaido University Press. Sapporo.
Webb, H. M. 1976. Interactions of daily and tidal rhythms. Pp. 129-
135 in Biological Rhythms in the Marine Environment. P. J. De-
Coursey, ed. University of South Carolina Press, Columbia.
Reference: Biol. Bull. 182: 265-269. (April, 1992)
Effects of Hypoxia and Anoxia on Larval Settlement,
Juvenile Growth, and Juvenile Survival
of the Oyster Crassostrea virginica
S. M. BAKER AND R. MANN
Virginia Institute of Marine Science, The College of William and Mary,
Gloucester Point, Virginia, 23062
Abstract. The effects of hypoxia (1.5 mg O: 1~', 20%
of air saturation) and anoxia (<0.07 mg O2 1~', <1% of
air saturation) on oyster (Crassostrea virginica) larval set-
tlement, juvenile growth, and juvenile survival were stud-
ied. Settlement was reduced significantly (P < 0.05) in
hypoxic treatments, as compared to normoxic treatments
(7.3 mg O2 h~', 100% of air saturation), and almost no
settlement took place in anoxic treatments. After 96 h,
38% and 4% of the larvae placed in hypoxic and anoxic
treatments had settled, while 79% settled in normoxic
treatments. In the first 144 h after settlement, juveniles
in hypoxic treatments grew one third as much as those
in normoxic treatments, while juveniles in anoxic treat-
ments did not grow at all. Median mortality times of re-
cently settled juveniles in hypoxic and anoxic treatments
were 131 h and 84 h, respectively. We conclude that hyp-
oxic and anoxic waters have potentially detrimental effects
on oyster settlement and recruitment.
Introduction
Chesapeake Bay exhibits episodes of oxygen depletion
concomitant with seasonal salinity and temperature strat-
ification (Taft et al., 1980; Officer el a/., 1984). Oxygen
depletion is usually restricted to areas below the pycno-
cline, but wind stress frequently tilts the pycnocline (Carter
el a!., 1978; Malone et a!., 1986) irrigating shallow areas,
where oyster reefs occur, with hypoxic or anoxic water
from deeper areas (May, 1973; Sanford et al., 1990). The
pycnocline remains tilted for from several hours to two
or three days (Malone et al.. 1986; Sanford et al.. 1987).
Received 26 August 1991; accepted 25 November 1991.
Contribution No. 1707 from the Virginia Institute of Marine Science,
School of Marine Science, The College of William and Mary.
These events often coincide with the timing of settlement
and recruitment of the oyster, Crassostrea virginica
Gmelin. Reduced settlement or complete settlement fail-
ure in localized areas has been attributed to incidents of
pycnocline tilting (May, 1973; Abbe, 1986).
Previous studies have demonstrated that tolerance of
larval and adult oysters to hypoxia and anoxia increases
with developmental stage and body size. Larval stages and
juvenile oysters (16 mm height) survive anoxia from hours
todays(Widdows?/ al.. 1989), while adult oysters survive
periods of unsuitable conditions lasting days or weeks
(Galtsoff, 1964; Stickle el al.. 1989).
Little is known about the tolerance of settling oyster
larvae or recently settled juvenile oysters to hypoxia and
anoxia. These stages are pivotal to subsequent recruitment
into the population. The objectives of this study, therefore,
were to examine the effects of hypoxia and anoxia on
settlement of oyster pediveliger larvae and on the growth
and survival of recently settled juvenile oysters.
Materials and Methods
Experimental apparatus
All experiments were performed at 25°C and 21%o S.
Temperature was maintained by controling laboratory
temperature and by a circulating water bath in which the
experimental chambers were immersed. Three 4-liter
flasks of 0.45 j/m filtered seawater containing algae (Iso-
chrysis galbana) at a concentration of 20,000 cells ml"1
were bubbled with air, a mixture of oxygen and nitrogen,
or nitrogen. The target oxygen concentrations were 7.3
mg O2 I'1 (100% of air saturation), 1.5 mg O2 1"' (20%
of air saturation), and less than 0.07 mg O2 I"1 (<1% of
air saturation). These treatments will be referred to as
265
266
S. M. BAKER AND R. MANN
normoxia, hypoxia, and anoxia, respectively, although the
latter of these conditions is more correctly termed 'mi-
croxia.' Although carbon dioxide was not included in the
latter two treatments, pH did not differ significantly (P
< 0.05, ANOVA) among the three treatments.
Flow-through chambers were constructed to hold larval
and juvenile oysters during experimental trials. Each
chamber was a 20 ml glass vial closed with a rubber stop-
per pierced by two 20 gauge needles. Inflow needles were
fitted with inverted pipette tips. Outflow needles were cut
off even with the bottom of the stoppers and covered with
202 ^m Nitex mesh, fine enough to retain pediveliger
larvae. Chambers within the same treatment were con-
nected in series as depicted in Figure 1. Stainless steel
tubing ( 1 mm bore) was used throughout. The flow rate
through the chambers was about 233 ml h~', and water
residence time in the system was 1 h or less. The flasks
of sea water and algae were replaced every 12 h with iden-
tical flasks that had been bubbled with the appropriate
gases for at least 2 h prior to replacement.
Oxygen concentration at the outflow of each treatment
was measured daily with a Strathkelvin Instruments (SI)
oxygen sensor ( 1 302) held in a SI microcell (MC 100) and
coupled to a SI oxygen meter (781) and chart recorder.
The oxygen sensor was calibrated daily with air-saturated
water and a 0% oxygen solution of sodium borate and
crystalline sodium sulfite. Normoxic, hypoxic, and anoxic
treatments were consistently maintained at 85-100%, 15-
22% and 0-1% of full air saturation, respectively. Outflow
concentrations of oxygen did not differ measurably from
the inflow concentrations.
Larval settlement experiments
Oyster (Crassostrea virginica) pediveliger larvae were
reared by the Virginia Institute of Marine Science oyster
hatchery at Gloucester Point, Virginia. Oyster shell set-
tlement substrates were conditioned in seawater for 24 h
prior to each experiment to develop a settlement-inducing
bacterial coating (Fitt et ai, 1990). One conditioned oyster
shell was placed in each chamber, with the rough side up.
Fifty larvae were counted into each chamber with a
Drummond Captrol III microdispensor. Only actively
swimming larvae were used.
Two chambers were removed daily from each treat-
ment; they were not replaced. Settlement was calculated
by expressing the number of settled oyster larvae as a
percentage of the total number of larvae introduced into
the chamber. The data from the two chambers were
pooled as one replicate for that exposure time. The entire
larval settlement experiment was repeated five times, re-
sulting in five replicates of normoxic treatments, and three
replicates each of hypoxic and anoxic treatments.
Larval settlement data were arcsine transformed, and
analysis of variance was performed for each exposure time
to test the null hypothesis that the means of the three
treatments were equal. For those exposure times in which
the null hypothesis was rejected, the Tukey multiple com-
parison test was performed to determine between which
treatment means differences existed (Zar, 1984). Means
and standard deviations were back transformed for report
in Figure 2.
Juvenile growth and sun'ival experiments
Unless otherwise noted, the term "juvenile" is used in
this paper to refer to those oysters 144 h post settlement
or less. Oyster pediveliger larvae were allowed to settle on
conditioned oyster shells for 2 h just prior to commence-
ment of the experiments. Non-settled larvae were washed
offafter 2 h. One oyster shell with settled larvae was placed
in each chamber, with the rough side up. Two chambers
were removed daily from each treatment; they were not
replaced. Twenty-five randomly selected live juvenile
oysters from each of the two chambers were measured
with a compound microscope and an ocular micrometer.
Growth was measured as the amount of new shell in the
dorsal-ventral axis (height). Mortality was recorded as the
proportion of dead juveniles among 50 randomly selected
juveniles from each chamber. The data from the two
chambers were pooled as one replicate for that exposure/
post settlement time. The entire juvenile growth and sur-
vival experiment was repeated four times, resulting in four
replicates of normoxic treatments, and three replicates
each of hypoxic and anoxic treatments.
Growth data were log transformed, and the residuals
were examined for homoscedasticity. Analysis of variance
was performed to test significance and linearity of the
growth regressions. Student's / test was used to determine
differences between the normoxic and hypoxic growth
regression coefficients and regression elevations (Zar,
1984).
Survival data for juvenile oysters were arcsine trans-
formed. Analysis of variance was performed for each
exposure/post settlement time to test the null hypothesis
that the means of the three treatments were equal. For
those exposure/post settlement times in which the null
hypothesis was rejected, the Tukey multiple comparison
test was performed to determine between which treatment
means differences existed (Zar, 1 984). Means and standard
deviations were back transformed for report in Figure 4.
Results
Larval settlement
In normoxic treatments at 24 h, the mean settlement
of oyster larvae was 38%. (Fig. 2). The percentage of settled
larvae increased 10-20% per day, and was 79% at 96 h.
In the hypoxic treatments, settlement was 18% at 24 h
WATER
RESERVOIR
EFFECTS OF HYPOXIA ON CRASSOSTREA
EXPERIMENTAL CHAMBER SERIES
Uow
267
AIRSTONE \
WATER
INTAKE
Figure 1. The experimental apparatus. Four chambers of one treatment are shown. Flasks of seawater
were bubbled with air, a mix of oxygen and nitrogen, or nitrogen. The equilibrated seawater was pumped
through chambers containing settlement substrate and pediveliger larvae or juveniles of the oyster Crassostrea
virginica. Flow-through chambers were immersed in a circulating water bath of 25°C. (Not drawn to scale.)
and 38% at 48 h. After 48 h, hypoxic treatments had no
further settlement. In anoxic treatments, settlement was
4% at 24 h, with no subsequent settlement. At 24 h, anoxic
and normoxic treatment means were significantly different
(P < 0.05), and at 48 h, the anoxic treatment mean was
significantly different (P < 0.05) from both the hypoxic
and normoxic treatment means. At 72 and 96 h, all three
treatment means were significantly different (P < 0.05)
from each other.
Juvenile growth
Regressions of log transformed juvenile oyster growth
data from normoxic and hypoxic treatments were linear
and significant. The regression coefficients of the normoxic
and hypoxic treatments were not significantly different;
however, the regression elevations were significantly dif-
ferent (P < 0.05) from each other (Fig. 3). Juveniles in
the normoxic treatments grew over 255 ^m of new shell
100i
Bl Anoxia p;%] Hypoxia | | Normoxia
80
v^
-L-
0)
•"-
E
.9 40-
i
•+r
ti
'
:>
&
W 20
,-
i
V
;'/
:
f '
v/
n
^
i
>
m
24 48 72 96
Hours exposure
Figure 2. Relation between percentage settlement of oyster (Cras-
sostrea virginica) pediveliger larvae and duration of normoxic (7.3 mg
O2 r1), hypoxic (1.5 mg O2 1~'), and anoxic (<0.07 mg O2 P1) treat-
ments. (Means + SD; normoxia n = 5; hypoxia n = 3; anoxia n = 3.)
o
O)
"55
D)
o
2
0
x — Normoxia
A Hypoxia
24 48 72 96 120 144
Hours post settlement
Figure 3. Log of growth of Crassostrea virginica juveniles (initial
shell height 290 ^m) in normoxic (7.3 mg O2 1~'), hypoxic (1.5 mg O2
1 " ' ). and anoxic ( <0.07 mg O2 1 ~ ' ) treatments in relation to hours post
settlement. (Means ± SD; normoxia n = 175 for each mean marker;
hypoxia n = 125 for each mean marker.)
268
S. M. BAKER AND R. MANN
in 144 h, nearly doubling in length. Juveniles in hypoxic
treatments grew only 77 jum of new shell in 144 h, ap-
proximately one third as much as those in normoxic
treatments. Juveniles in anoxic treatments did not increase
in shell height.
Juvenile survival
Juvenile oyster survival was similar in all three treat-
ments for the first 72 h (Fig. 4). At 96 h and 120 h, the
anoxic treatment mean was significantly different (P
< 0.05) from both hypoxic and normoxic treatment
means. All three treatment means were significantly dif-
ferent (P < 0.05) from each other at 144 h. Juveniles in
the anoxic treatments had a median mortality time (time
to 50% mortality) of 84 h. Mortality of juveniles in anoxic
treatments was 100% by 144 h. Juveniles in the hypoxic
treatments had a median mortality time of 1 3 1 h. Nor-
moxic treatments, in contrast, had a mean of only 13%
mortality at 144 h.
Discussion
Under hypoxic and anoxic conditions, oyster pediveli-
ger larvae significantly reduce energetically costly activi-
ties, thereby reducing total metabolism and oxygen re-
quirements (Widdows el a!., 1989). The results of this
study indicate that settlement is another costly activity
that oyster pediveliger larvae avoid when in oxygen-lim-
iting environments.
In a recent paper on the effects of hypoxia and anoxia
on Mytilus edulis larvae, Wang and Widdows (1991) re-
port that moderate hypoxia has little effect on larval set-
tlement. Settlement of mussel pediveliger larvae onto adult
byssus filaments is approximately 1 2% after two days in
conditions of 8.2 mg O2 P' (20.0 kPa pO2, 98% of air
saturation at 15°C and 31%»), 2.4 mg O2 I"1 (5.91 kPa
pO2, 29% of air saturation), or 1.3 mg O2 1~' (3.16 kPa
pO2, 15% of air saturation). An oxygen concentration of
0.6 mgO: P1 (1.38 kPa^O2. 7% of air saturation) shows
1% settlement. Settlement of C. virginica appears to be
more sensitive to moderate hypoxia than mussel settle-
ment. While settlement of mussel larvae is unchanged in
treatments of 8. 2 mgO2 1~' down to 1.3mgO2 P1 (Wang
and Widdows, 1991), oyster larval settlement was signif-
icantly reduced by oxygen concentrations of 1.5 mg O2
1~' or less. The estimated oxygen concentration at which
settlement after two days is 50% of that in normoxic treat-
ments is 0.9 mg O2 1~' (10% of air saturation) for mussel
larvae (Wang and Widdows, 1991) compared to 1.4 mg
O2 P1 (20% of air saturation) for oyster larvae. While
oysters are entirely sessile once they have settled, post
larval mussels migrate repeatedly before arriving at a final
settlement site (Lane et al, 1985). Larval mussels, there-
Anoxia
100
03
t±
O
E
E
3
O
0 24 48 72 96 120 144
Hours exposure/post settlement
Figure 4. Relation between cumulative mortality of Crassostrea vir-
gmica juveniles and duration of normoxic (7.3 mgO2 1~'), hypoxic (1.5
mg Oi r'), and anoxic (<0.07 mg O2 1~') treatments. Arrows indicate
median mortality times. Where no standard deviation is shown, the
standard deviation is smaller than the mean marker. (Means ± SD; nor-
moxia n = 4; hypoxia n = 3; anoxia n = 3)
fore, do not need to be as discriminating as oyster larvae
when selecting a suitable settlement habitat.
In other aspects of their physiology, oyster larvae are
less sensitive to oxygen deprivation than are mussel larvae.
For example, the oxygen concentration at which the res-
piration rate is 50% of the normoxic rate is 2.3 mgO2 P'
(5.7 kPa^O2, 28% of air saturation) for mussel pediveliger
larvae (Wang and Widdows, 1991) and 0.9 mg O2 I"1
(2.3 kPa pO2, 11% of air saturation at 22°C and 12%o)
for oyster pediveliger larvae (Widdows et al, 1989). The
10°C difference in temperature at which the mussel (Wang
and Widdows, 1991) and oyster (this paper) settlement
experiments were performed, and the resulting differences
in metabolic rates, may have contributed to the discrep-
ancy observed in oxygen sensitivity of mussel and oyster
larval settlement. At 15°C, mussel pediveliger larvae have
a normoxic oxygen uptake of 75 pmol O2 h~' larva"1
(Wang and Widdows. 1991), while at 22°C, oyster pedi-
veliger larvae have an oxygen uptake of 400 pmol O2 IP1
larva"1 (Widdows el al.. 1989).
As discussed earlier, pediveliger larvae reduce energet-
ically costly activities during hypoxic exposure, such as
ingestion, digestion, and growth, thereby reducing oxygen
demand. Under hypoxic conditions, there is a marked
decline in the proportion of pediveliger larvae feeding and
in ingestion rates (Widdows et al., 1989). Mussel pedi-
veliger larvae also exhibit depressed feeding rates and
growth in hypoxic conditions (Wang and Widdows, 1 99 1 ).
The reduction of juvenile oyster growth in hypoxic treat-
ments and complete lack of growth in anoxic treatments
EFFECTS OF HYPOX1A ON CR.4SSOSTREA
269
observed in this study may have resulted from a cessation
of feeding.
In this study, juvenile oysters had a median mortality
time of 84 h in anoxia. This indicates that, like oyster
larvae and adults, recently settled juvenile oysters are ca-
pable of anaerobic metabolism. Widdows el al. (1989)
report median mortality times in anoxia of 1 1, 18, and
5 1 h for oyster prodissoconch, veliconch, and pediveliger
larvae, and 150 h for juveniles 16 mm in shell height.
The data for recently settled juveniles are consistent with
the trend of increasing anoxic tolerance with develop-
mental stage and body size. The increased median survival
time in later stages is associated with an ability to reduce
energy use, measured as heat dissipation, under anoxic
conditions (Widdows el al.. 1989). The degree of heat
dissipation reduction by recently settled juvenile oysters
in anoxia is expected to be between that of the pediveliger
larvae and 16 mm juveniles studied by Widdows et al.
(1989).
Further studies on feeding, heat dissipation, and oxygen
uptake are required to understand more clearly the effects
of anoxia and hypoxia on settling pediveliger larvae and
recently settled juvenile oysters. The present study does
demonstrate that hypoxic and anoxic conditions have
detrimental effects on larval settlement, juvenile growth,
and juvenile survival. Oyster distribution may be influ-
enced by anoxia and hypoxia, especially in those areas
that experience prolonged (longer than 48 to 72 h) or
severe (anoxic) pycnocline tilt events. Pycnocline tilt
events may control recruitment into the adult population
directly, because of larval settlement failure and juvenile
mortality, and indirectly, because of a reduction in the
growth rate of juveniles.
Acknowledgments
This study was supported by funds from the National
Oceanic and Atmospheric Administration to RM and
the International Women's Fishing Association to SMB.
We thank the staff of the VIMS oyster hatchery for the
provision of larvae. P. Baker, B. Barber, L. Schaffner,
R. I. E. Newell, and two anonymous reviewers made
helpful comments on the manuscript.
Literature Cited
Abbe, G. R. 1986. A review of some factors that limit oyster recruitment
in Chesapeake Bay. Am Malacol. Bull.. Special Edition No. 3: 59-
70.
Carter, H. H., R. J. Regier, E. W. Schiemer, and J. A. Michael.
1978. The summertime vertical distribution of dissolved oxygen at
the Calvert Cliffs generating station: a physical interpretation. Ches-
apeake Bay Institute, The Johns Hopkins University, Special Report
60: 1-95.
Fitt, W. K., S. L. Coon, M. Watch, R. M. Weiner, R. R. Colwell, and
D. B. Bonar. 1990. Settlement behavior and metamorphosis of
oyster larvae (Crassostrea gigas) in response to bacterial supernatants.
Mar. Biol 106: 389-394.
Galtsoff, P. S. 1964. The American Oyster Crassostrea virginica Gme-
lin. Fish. Bull. Fish Wildlife Ser. 64: 1-480.
Lane, D. J. W., A. R. Beaumont, and J. R. Hunter. 1985. Byssus drifting
and the drifting threads of the young post-larval mussel Mytilus edulis.
Mar Biol 84: 301-308.
Malone, T. C, W. M. Kemp, H. W. Ducklow, W. R. Boynton, J. H.
Turtle, and R. B. Jonas. 1986. Lateral variation in the production
and fate of phytoplankton in a partially stratified estuary. Mar. Ecol.
Prog. Ser. 32: 149-160.
May, E. B. 1973. Extensive oxygen depletion in Mobile Bay, Alabama.
Limnol Oceanogr. 18(3): 353-366.
Officer, C. B., R. B. Biggs, J. L. Taft, L. E. Cronin, M. A. Tyler, and
W. R. Boynton. 1984. Chesapeake Bay anoxia: origin, development,
and significance. Science 223: 22-27.
Sanford, L. P., K. G. Sellner, and D. L. Breitburg. 1990. Covariability
of dissolved oxygen with physical processes in the summertime Ches-
apeake Bay. J. Mar. Res. 48: 567-590.
Sanford, L., K. Sellner, and M. Bundy. 1987. Moored measurements
of dissolved oxygen in the Chesapeake Bay during the summer of
1987. AGU Ocean Sciences Meeting, New Orleans, LA, 1987.
Stickle, W. B., M. A. Kapper, L.-L. Liu, E. Gnaiger, and S. Y. Wang.
1989. Metabolic adaptations of several species of crustaceans and
molluscs to hypoxia: tolerance and microcalorimetric studies. Biol
Bull. 177:303-312.
Taft, J. L., E. D. Hartwig, and R. Loftus. 1980. Seasonal oxygen de-
pletion in Chesapeake Bay. Estuaries 3(4): 242-247.
Wang, W. X., and J. Widdows. 1991. Physiological responses of mussel
larvae Mytilus edulis to environmental hypoxia and anoxia. Mar.
Ecol Prog. Ser. 70: 223-236.
Widdows, J., R. I. E. Newell, and R. Mann. 1989. Effects of hypoxia
and anoxia on survival, energy metabolism, and feeding of oyster
larvae (Crassostrea virginica, Gmelin). Biol. Bull. 177: 154-166.
Zar, J. H. 1984. Biostatistical Analysis. Prentice-Hall, Englewood Cliffs.
New Jersey.
Reference: Biol. Bull. 182: 270-277. (April, 1992)
Developmental Changes in Ionic and Osmotic
Regulation in the Dungeness Crab, Cancer magister
A. CHRISTINE BROWN' AND NORA B. TERWILLIGER
Oregon Institute of Marine Biology, University of Oregon, Charleston, Oregon 97420, and
Department of Biology, University of Oregon, Eugene. Oregon 97403
Abstract. The ontogeny of osmoregulation and specific
ion regulation was studied in the megalopa, 1st instar ju-
venile, 5th instar juvenile and adult of Cancer magister.
Hemolymph Na+, Cl~, K+, Mg++, and Ca++ concentra-
tions and osmolality were measured after 8-h exposure to
100%, 75%, and 50% seawater at 10°C and 20°C. The
ability to hyperosmotically regulate is present in the
megalopa, and ontogenic changes occur in both ionic and
osmotic regulation. First instar juvenile crabs, which are
exposed to the greatest extremes of salinity and temper-
ature in the field, are less able to osmoregulate than are
the other three stages examined. Changes in Na+, Cl~,
and K+ concentrations parallel total osmolality in all four
stages. Hemolymph Mg++ concentrations in megalopa
and juveniles acclimated to 100% seawater are more than
twice that of the concentration in the adult; after 8 h in
50% seawater, the megalopa and juvenile Mg++ concen-
trations decrease to the level of the strongly regulated adult
Mg++ concentration. Ca++ is strongly regulated by mega-
lopas and adult crabs exposed to reduced salinity com-
pared to the two juvenile stages. Diminished predation
pressure and high food availability are proximate factors
that may outweigh short-term osmoregulatory stress en-
countered on the tideflats during development of the ju-
venile crab.
Introduction
Estuarine invertebrates vary greatly in their abilities to
deal with changes in ambient salinity. The effects of en-
vironmental salinity on the internal osmolality and spe-
cific ion regulation of adult estuarine crustaceans have
Received 26 August 1991; accepted 21 January 1992.
1 Present Address: Department of Biology. Lake Forest College, Lake
Forest. IL 60045.
been investigated in numerous studies (for review, see
Mantel and Farmer, 1983). Ontogeny of osmoregulation
and ion regulation has been comprehensively studied in
branchiopod crustaceans, especially the anostracan brine
shrimp, Anemia (for review, see Conte, 1984). Compa-
rable information about larval, post-larval, and juvenile
decapod crustacean osmoregulation is relatively limited
(Kalber, 1970; Foskett, 1977; Young, 1979; Read, 1984;
Rabalais and Cameron, 1985; Charmantier el ai, 1988;
Charmantier and Charmantier- Daures, 1991), and there
are almost no data available regarding specific ion regu-
lation during decapod crustacean development (Char-
mantier et al, 1984a,b,c; Feldere/a/.. 1986).
The Dungeness crab, Cancer magister, inhabits the cold
waters of the Pacific Northwest coast of North America
and uses different portions of the estuarine and nearshore
waters during its life cycle. Along the Oregon coast, em-
bryos hatch from December through March (Reed, 1969;
Lough, 1976). The newly hatched larvae go through five
zoeal stages, all of which are planktonic in ocean waters,
moving as far as 200 miles offshore. The transitional stage,
an actively swimming planktonic megalopa, reenters the
coastal and estuarine waters from mid April through early
July (Lough, 1976). The megalopas then metamorphose
into 1st instar juveniles that join the benthic community.
Throughout the summer, juvenile crabs in the estuary are
found in high numbers on the tideflats, while the adult
crabs occur mainly in the deeper channels. Summer tidal
changes in salinity and temperature, extending over a pe-
riod of 6-8 h, are much greater on the tideflats than in
the estuarine channels. Adults of C. magister do not mi-
grate up into brackish waters for long periods as does Cal-
linectes sapidus, the East and Gulf coast blue crab, but
remain in the lower half of the bay, moving back and
forth into nearshore waters.
270
CRUSTACEAN ION REGULATION ONTOGENY
271
Previous studies on osmotic and ionic regulation in
Cancer magister reported that adult crabs were weak hy-
perosmoregulators after 72-96 h exposure to dilute con-
centrations of seawater (Jones, 1941; Alspach, 1972; En-
gelhardt and Dehnel, 1973; Hunter and Rudy, 1975).
These authors found that in reduced salinity, adult crabs
were able to strongly hyporegulate magnesium and hy-
perregulate calcium. Hemolymph sodium, chloride, and
potassium were hyperregulated but not as strongly as cal-
cium. None of these studies examined the responses of
larval, megalopa, or juvenile C. magister to changes in
salinity.
In this paper, we ask how adults of C. magister respond
osmotically to ecologically relevant short term tidal cycle
changes in salinity. We also investigate whether the adult
osmotic response pattern is present in the megalopa and
juvenile stages or whether there is an ontogeny of osmotic
regulatory abilities. Finally, we compare specific ion reg-
ulation during these life stages to see whether regulatory
abilities for individual ions occur differentially during on-
togeny.
Materials and Methods
Animals
Megalopas of Cancer magister (Dana) were collected
with a dip net from the surface waters of Coos Bay, Or-
egon, from April through June of 1989. Because the
megalopas molt within 48-72 h after collection, they were
used in experiments within two days. In the laboratory,
megalopas were kept in glass and wood aquaria (10 gal-
lons) with running aerated seawater pumped on an in-
coming tide from near the mouth of Coos Bay. Salinity
was 30-33%o, and water temperature was 10-12°C.
Megalopas were not fed.
Juvenile crabs were raised from field-caught megalopas
and were maintained in similar aquaria with running sea-
water and aeration. Adult males of C. magister, collected
from the Coos Bay channel using crab pots, were kept in
large holding tanks (260 gallons) with running seawater
and aeration at the same temperature and salinity as
megalopas and juveniles. Both juveniles and adults were
fed 3-5 times a week on mussels, fish, and squid. Feeding
was stopped 24 h prior to experiments to ensure a post-
absorptive state in the crabs and to avoid their fouling the
experimental chamber.
Protocol and sampling
Experiments were run on intermolt animals with ju-
venile intermolt stage based on time elapsed since the
preceding molt. Thus, intermolt 1st instar juveniles were
available in April-June, 5th instar juveniles in September-
November, and adults in December-February.
Megalopas (approx. 3 mm carapace width), 1st instar
juveniles (6-8 mm carapace width), 5th instar juveniles
(25-33 mm carapace width), and adults (larger than 120
mm carapace width) were exposed to test conditions at
varied temperatures and salinities for a period of 8 h.
Hemolymph samples were taken immediately thereafter
for osmotic and ionic analyses. Test conditions included
100% seawater (32%o, obtained on an incoming tide at
the mouth of Coos Bay), 75% seawater, and 50% seawater
(Coos Bay seawater diluted with glass distilled water)
maintained at both 10°C and 20°C. Glass aquaria (one
gallon) were used for the experimental chambers. About
250 megalopas or 1st instar juveniles and 2 or 3 5th instar
juveniles were placed in each aquarium. Adults were kept
one to an aquarium for the duration of the experiments.
Hemolymph was taken from the megalopas by punc-
turing the heart with a glass micro-capillary pipette. Ju-
veniles and adults were bled by puncturing the arthrodial
membrane at the base of a walking leg; 1st instar juveniles
were bled with micro-capillary pipettes, 5th instar juve-
niles and adults were bled with needle and syringe. He-
molymph obtained from all individuals in each experi-
mental aquarium was pooled in order to collect a single
sample of sufficient volume for both osmotic and ionic
analyses. In figure legends 1-6, n refers to the number of
separate pooled samples on which analyses were per-
formed (megalopa, n = 1-3; 1st and 5th instar juveniles,
n = 2-3; adult, n = 8). Seawater samples from each
aquarium were collected. Samples were immediately fro-
zen and stored at -73°C for subsequent osmotic and ionic
analyses.
Osmotic and ionic analyses
Osmolality of seawater and hemolymph samples was
measured using a Wescor 5500 vapor pressure osmometer.
Chloride concentration was measured using a Buchler-
Cotlove chloridometer. Magnesium concentration was
measured colori metrically after the method of Sky-Peck
(1964). That is, samples were deproteinized with 5% tri-
chloroacetic acid and reacted with thiazole yellow in the
presence of excess base. The absorbance at 540 nm was
measured with a Beckman DU-70 spectrophotometer.
Sodium, calcium, and potassium ion activities were mea-
sured with a Radiometer Ion 83 ion meter in mV mode
and the following electrodes: Radiometer G502 sodium
Selectrode, Microelectrodes Inc. MI-420 sodium micro-
electrode. Radiometer F2112 calcium Selectrode, and
Orion 90- 1 9 potassium electrode. The reference electrode
in all cases was an Orion 90-02 double junction reference
electrode with an NH4C1 outer chamber filling solution
and a AgCl saturated inner chamber filling solution. Sam-
ples were diluted 1:100 in the appropriate ionic strength
adjustment solution. Prior to the analysis of samples, cal-
272
A. C. BROWN AND N. B. TERWILLIGER
ibration for measurement of each ion species was done
with salt solutions of known concentration spanning the
expected range of values.
Data analysis
Results are expressed as mean ± S.E. (n = number of
observations). Three-way analysis of variance (ANOVA)
was used to test for significance among treatments (de-
velopmental stage, salinity, and temperature). Subsequent
multiple comparisons of means were performed using the
Tukey-Kramer method. Statistical significance was ac-
cepted at P < 0.05.
Results
Estuarine salinity and temperature were measured in
areas where the different developmental stages of C. mag-
ister were abundant in order to set limits for these param-
eters in laboratory studies. The tideflat environment of
juveniles ranges from 10°C at high tide to 25 °C when the
tide has receded and tideflats are exposed during early-
to mid-morning low tides in summer. At the same time,
salinity drops from 32 to 16%o as the freshwater lens on
the surface passes down the flats. In the channels where
adults are found, summer water temperature (10-15°C)
and salinity (32-20%o) are much more stable. Winter water
temperature is consistently low ( 10-12°C). Winter range
of salinity (32-16%o) at depth in the estuary varies as
widely as salinity on the summer tideflats owing to the
increased fresh water input from rain.
Osmoregulation
After 8-h exposure to 100% seawater, the megalopa,
1st instar juvenile, 5th instar juvenile, and adult are isos-
motic with the ambient seawater (Fig. 1 ). In 75% seawater,
the hemolymph osmolalities of all four stages are signif-
icantly lower than in 100% seawater, yet they are all hy-
perosmotic relative to 75% seawater. In 50% seawater, the
hemolymph osmolalities of all four stages are significantly
lower than in 75% seawater, and all are significantly hy-
perosmotic compared with 50% seawater. The 1st juvenile
is least able to maintain hemolymph osmolality in dilute
seawater compared to the other stages examined. The
crabs are less able to osmoregulate in warmer water. He-
molymph osmolalities of the adult and 5th instar juvenile
are significantly lower at 20°C than at 10°C in both 75%
and 50% seawater; megalopa hemolymph osmolality is
also lower at 20°C than at 10°C in 50% seawater.
Ionic regulation
The hemolymph chloride concentration in all four
stages in 1 00% seawater is hypoionic compared with am-
bient seawater (Fig. 2). In 75% seawater the adult becomes
1000
o
£ 400
500-
1000
900-
800
700
600
500
20°C
400
400
500 600 700 800 900
Medium Osmolality (mOsm/kg)
1000
Figure 1 . Hemolymph osmolality of Cancer magister as a function
of medium osmolality for *, megalopa (n = 1-3); A, 1st instar juvenile
(n = 2-3); •, 5th instar juvenile (n = 2-3); • adult (n = 8). Solid
symbols, n > 2, standard error bars drawn; open symbols, n < 2, mean.
nearly isoionic compared with the seawater and has sig-
nificantly lower hemolymph chloride concentration at
20°C than at 10°C. In 50% seawater the adult hemolymph
chloride concentration is hyperionic and is lower at 20°C
than at 10°C. The hemolymph chloride concentrations
of the megalopa and of the 5th instar juvenile are also
temperature sensitive in 75% seawater. In 50% seawater
the megalopa and 1st instar juvenile hemolymph chloride
concentrations are the same as the ambient seawater
chloride, while that of the 5th instar juvenile is significantly
higher than that of the megalopa and 1st instar juvenile
and lower than that of the adult.
Hemolymph sodium ion activity (Fig. 3) in all four
stages shows essentially the same pattern as hemolymph
chloride. In 100% seawater all four stages are hypoionic
with respect to ambient seawater sodium. In 75% seawater
the megalopa, 1st instar juvenile, and 5th instar juvenile
hemolymph sodium ion activities are significantly less
than in 100% seawater, while the adult hemolymph so-
dium ion activity is not significantly changed. In 50% sea-
water the megalopa and 1st instar juvenile hemolymphs
are isoionic to ambient sodium, whereas the 5th instar
juvenile hemolymph sodium is intermediate between the
younger stages and the adult. There is no significant effect
CRUSTACEAN ION REGULATION ONTOGENY
273
550
o
E
= 250-
300-
250
200
200 250 300 350 400 450 500
Medium chloride ion concentration (mmol/L)
550
Figure 2. Hemolymph chloride ion concentration of Cancer magisier
as a function of medium chloride ion concentration for 4, megalopa (n
= 1-3); A, 1st instar juvenile (n = 2-3); •. 5th instar juvenile (n = 2-
3); • adult (n = 8). Solid symbols, n > 2, standard error bars drawn;
open symbols, n < 2, mean.
of temperature on the hemolymph sodium ion activity in
any of the stages.
There is no significant effect of temperature on he-
molymph potassium ion activity (Fig. 4). The megalopa
shows no significant change in hemolymph potassium ion
activity. The adult, 1st instar juvenile, and 5th instar ju-
venile hemolymph potassium ion activities, however, are
significantly less in 75% and 50% seawater than in 100%
seawater.
In contrast to the concentrations of chloride, sodium,
and potassium, that of magnesium is strongly hyporegu-
lated in adult hemolymph in all salinity treatments (Fig.
5). In 100% seawater the megalopa, 1st instar juvenile,
and 5th instar juvenile hemolymph magnesium concen-
trations are significantly higher than the adult. As salinity
decreases, magnesium concentration in these three stages
also decreases until, in 50% seawater, there is no difference
in the hemolymph magnesium concentration among all
four stages. The only stage in which magnesium regulation
shows a significant temperature sensitivity is the 5th instar
juvenile in 100% seawater.
In 100% seawater the hemolymph calcium ion activities
in all four stages are not significantly different from the
ambient seawater calcium ion activity (Fig. 6). The he-
molymph calcium ion activities of the megalopa and adult
do not change significantly with salinity. The 1st instar
juvenile and 5th instar juvenile, however, have signifi-
cantly lower hemolymph calcium ion activities in 75%
and 50% seawater than in 100% seawater. Overall there
is no significant effect of temperature on hemolymph cal-
cium ion activity.
Discussion
Different developmental stages in the life cycle of Can-
cer magister have distinctly different patterns of hemo-
lymph osmotic and ionic regulation when exposed to re-
duced salinity. The values for hemolymph osmolality and
ionic concentrations in the present study were obtained
after an 8-h exposure time, the duration of a tidal cycle,
which is physiologically and ecologically relevant for these
crabs. Furthermore, the general trends for osmotic and
ionic regulation reported in the long-term 72-96 h equi-
librium exposures (Jones, 1941; Alspach. 1972; Engelhardt
and Dehnel, 1973; Hunter and Rudy, 1975) are apparent
after the 8-h exposure time used here. We find, just as the
450
200 250 300 350 400
Medium sodium ion activity (mmol/L)
450
Figure 3. Hemolymph sodium ion activity of Cancer magister as a
function of medium sodium ion activity for 4, megalopa (n = 1-3); A.
1st instar juvenile (n = 2-3); •, 5th instar juvenile (n = 2-3); • adult,
(n = 8). Solid symbols, n > 2, standard error bars drawn; open symbols,
n < 2, mean.
274
A. C. BROWN AND N. B. TERW1LLIGER
a.
>*
o
123456
Medium potassium ion activity (mmol/L)
Figure 4. Hemolymph potassium ion activity of Cancer magister as
a function of medium potassium ion activity for #, megalopa (n = 1-
3); A, 1st instar juvenile (n = 2-3); •, 5th instar juvenile (n = 2-3); •
adult (n = 8). Solid symbols, n > 2, standard error bars drawn; open
symbols, n < 2, mean.
earlier equilibrium studies reported, that the hemolymph
of adult C. magister is weakly hyperosmoregulated in wa-
ter less concentrated than normal ocean seawater; chlo-
ride, sodium, and potassium are somewhat hyperregulated
in reduced salinity, magnesium is very strongly hyporeg-
ulated and calcium is strongly hyperregulated. Compared
with adults of four other species within the genus Cancer
for which data are available (see Charmantier and Char-
mantier-Daures, 1991, for review), C. magister adults are
the strongest osmoregulators. For example, Cancer an-
tennahus has a hemolymph osmolality only 1 5 mOsm/
kg above ambient seawater osmolality in approximately
53% seawater at 15-20°C (Jones, 1941) compared with
C. magister hemolymph osmolality of 250 mOsm/kg
above ambient seawater in 50% seawater at 20°C.
Ontogeny of osmoregulation
Studies on the larvae, post-larvae, and juveniles of a
number of decapod crustacean species indicate that most
larvae and post-larvae can maintain hemolymph osmo-
lality above that of ambient seawater, either by hyperos-
moconforming or by weakly hyperosmoregulating (see
Charmantier et al., 1988). In such cases, metamorphosis
often marks a profound change in osmoregulation from
larval to adult patterns. Many decapods that are hyperos-
moconforming or weakly hyperosmoregulating over a
wide range of salinities in the premetamorphic stages un-
dergo a change to become either (a) strongly hyperos-
moregulating in low salinity and osmoconforming in high
salinity in the adult, as in Clibanarius vittatus, Homarus
gammanis. Homarus atnericanus, and Cancer irroratus
(Young, 1979;ThuetfM/., 1988; Charmantier et al. 1988;
Charmantier and Charmantier-Daures, 1 99 1 ) or (b) hy-
per-hypoosmoregulating in the adult, as in Sesarrna re-
ticulatum, Uca subcylindricum, and Penaeus japonicus
(Foskett. 1977; Rabalais and Cameron, 1985; Charman-
tier et ai, 1988). Less frequently described are species in
which the larvae have a greater osmoregulatory ability
than the adult, i.e.. Hepatus epheliticus and Libinia
emarginata, but a switch still occurs around the time of
metamorphosis (Kalber, 1970). Macrobrachium pete rsi is
a case in which both larvae and adults are strong osmo-
regulators, but the different stages vary in their capacity
to hypo or hyperosmoregulate (Read, 1984). To date the
10
10
20 30 40 50
Medium magnesium ion concentration (mmol/L)
Figure 5. Hemolymph magnesium ion concentration of Cancer
magister as a function of medium magnesium ion concentration for *,
megalopa (n = 2); A, 1st instar juvenile (n = 2); •, 5th instar juvenile
(n = 2-3); « adult (n = 8). Solid symbols, n > 2. standard error bars
drawn; open symbols, n < 2. mean.
CRUSTACEAN ION REGULATION ONTOGENY
275
Q.
J>*
O
56789
Medium calcium ion activity (mmol/L)
Figure 6. Hemolymph calcium ion activity of Cancer magister as a
function of medium calcium ion activity for $, megalopa (n = 1-3); A,
1st instar juvenile (n = 2-3); •, 5th instar juvenile (n = 2-3); • adult
(n = 8). Solid symbols, n> 2, standard error bars drawn; open symbols,
n < 2, mean.
only decapod species whose osmoregulatory pattern does
not change during development is Callianassa jamaicense
var. louisianensis. Its larvae remain in the hyposaline
burrow habitat of the adult, and all stages show limited
hyperosmoregulation in dilute media, although the adult
can hyperregulate a bit more strongly than the zoeae
(Felder, 1978; Felder et al., 1986).
In the present study, there is a marked change in os-
moregulation at metamorphosis from megalopa to 1st in-
star juvenile in C. magister. Interestingly, the juvenile C.
magister is less able to regulate over the short 8-h exposure
than is the megalopa (Fig. 1). As development proceeds,
however, osmoregulatory ability becomes more like that
of the adult. A correlation between ontogeny of osmo-
regulation and changes in habitat salinity has been ob-
served for several species (for review, see Charmantier et
al., 1988). This correlation does not hold as an explanation
for the observed patterns of osmoregulation in the different
stages of C. magister. The 1st instar juvenile, the stage
least able to osmoregulate, is found in high numbers on
the mudflats where it encounters extremes of low salinity
and high temperature. Other factors, such as the short
duration of exposure to these extremes and behavioral
responses, must play a role in environmental distribution.
Both juvenile and adult C. magister become inactive in
low salinities. This behavior may enable the juveniles to
endure the low salinity portion of the tide cycle on the
mudflat without great expenditure of energy while expe-
riencing protection from heavy predation by adult crabs
and fish concentrated during low tide in the deeper chan-
nels. As the tide rises, juveniles on the mudflats are able
to immediately resume active foraging; proximity to high
food availability may offset the short term osmoregulatory
stresses experienced during low tide.
Several factors may be responsible for the diminished
ability of the 1st instar juvenile to osmoregulate. Although
megalopas, 1st, and 5th instar juveniles all have gills that
function in ion transport based on silver staining (Brown,
unpub. obs.), there are no data available on the ratio of
gill surface area to total body volume in these stages. The
carapace of the 1st instar juvenile is about twice as wide
as in the megalopa, and the juvenile weighs twice as much;
if the gill surface area has not increased proportionately,
this might explain the diminished osmoregulatory capacity
of the juveniles. Also important are the efficiency of salt
transport at the gill and the amount of area on the gill
associated with that salt transport (see Conte, 1984). Felder
et al. (1986) have shown differences in Na+/K+ ATPase
activity in the different prehatch stages of Callianassa ja-
maicense var. louisianensis and have demonstrated the
presence of salt transport type tissue on the brancheo-
stegites of the zoeae. Homarus gammarns post-larvae,
which have greater osmoregulatory capacities compared
with larval stages, show a marked increase in gill Na+/K+
ATPase and carbonic anhydrase activities (Thuet et al.,
1988). It is possible that the juvenile stages of C. magister
initially have lower Na+/K+ ATPase activity levels or a
different relative proportion of salt transporting tissue than
the megalopas.
Ontogeny of ion regulation
This is the first report of an ontogenic change in specific
ion regulation in brachyuran crabs. The data show that
specific ions are regulated differently by megalopa, juve-
nile, and adult C. magister. At metamorphosis from
megalopa to 1st instar juvenile, there are shifts in specific
ion regulatory patterns that do not parallel changes in
hemolymph osmolality. Ion regulation in 5th instar ju-
venile is more like that in adult than that in 1st instar
juvenile; the fifth instar juvenile is beginning to show the
adult pattern of ion regulation but does not regulate to
the same extent as the adult.
Hemolymph levels of sodium, chloride, and potassium
parallel stage-specific changes in hemolymph osmolality
as salinity decreases. Sodium and chloride are the two
276
A. C. BROWN AND N. B. TERWILLIGER
main inorganic ion constituents in the hemolymph and
appear to be the major components in osmotic regulation
in the different developmental stages.
Two aspects of developmental changes in specific ion
regulation are particularly striking. First, hemolymph
magnesium concentrations in megalopas and in 1st and
5th insta juveniles in 100% seawater are twice as high as
in the adult, when none of the other ions show any dif-
ferences between stages in 100% seawater. Second, calcium
is strongly hyperregulated in megalopa and adult hemo-
lymph, as salinity decreases, compared to the two juvenile
stages studied.
Adults of all species of crustaceans that have been ex-
amined maintain hemolymph magnesium well below the
magnesium concentration of the ambient water, except
when they are in extremely dilute water. Engelhardt and
Dehnel (1973) stated that "hyporegulation of magnesium
is the most universal feature of ionic regulation in crus-
tacean blood." Adult C. magister excrete magnesium in
urine formed in the antennal gland; the urine to hemo-
lymph ratio of magnesium is nearly 4:1 in 100% seawater
(Hunter and Rudy, 1975; Holliday, 1980). In the early
stages of crustaceans, the antennal gland may not be fully
developed and functional (Waite, 1899; Conte, 1984), and
this may account for the high hemolymph magnesium in
megalopa and juvenile crabs. Low hemolymph magne-
sium levels have often been associated with high levels of
activity or a greater extent of terrestriality in crustaceans.
According to Robertson (1960), decapod species with he-
molymph magnesium concentrations less than 50% that
of seawater are more active than those with higher he-
molymph magnesium concentrations. In fact, high mag-
nesium concentrations are often used to anaesthetize ma-
rine invertebrates. The totally aquatic C. magister mega-
lopa is an extremely active animal, however, capable of
swimming very rapidly for extended periods. The 1st instar
juvenile, which like the megalopa has more than twice
the adult's hemolymph magnesium, is also considerably
more active than the adult. Gross ( 1964) discusses mag-
nesium regulation at length in relation to the extent of
terrestriality of various crab species. Mantel and Farmer
(1983) note that grapsids and other species of semi-ter-
restrial and terrestrial decapods all have low hemolymph
magnesium concentrations. Because all stages in the life
cycle of C. magister are aquatic, the change from higher
to lower magnesium concentration we observe during de-
velopment is not related to changes in extent of terrestri-
ality. In summary, the high magnesium in the hemolymph
of the megalopa and juvenile and the developmental
changes in magnesium levels are consistent with neither
of these hypotheses, activity level or terrestriality.
Calcium regulation in C. magister adults has been pre-
viously reported (Alspach, 1972; Engelhardt and Dehnel,
1973; Hunter and Rudy, 1975). The strong regulation of
calcium by the megalopa and weak regulation by the ju-
veniles is noteworthy. The low levels of calcium in the
hyposaline-exposed 1st and 5th instar juveniles may reflect
their overall decreased ability to regulate ions.
In summary, tidal cycle changes in salinity and tem-
perature have a strong effect on hemolymph osmolality
and ionic concentration in megalopas, juveniles, and
adults of C. magister. There are ontogenic changes in both
ionic and osmotic regulation in C. magister. Calcium and
magnesium regulation change markedly during devel-
opment from megalopa to adult crab. Both of these ions
have strong effects on the oxygen affinity and the coop-
erativity of hemocyanin from a variety of crustacean spe-
cies (Larimer and Riggs, 1964; Miller and Van Holde,
1974; Truchot, 1975). The changes in hemolymph cal-
cium and magnesium levels during the development of
C. magister may be involved in modulating the oxygen-
binding properties of the hemocyanin. This hypothesis is
currently under investigation.
Acknowledgments
This study was supported by NSF DMB 85-1 1 150 and
DCB 89-08362 (NBT) and the Lerner-Grey Fund for Ma-
rine Research (ACB). This is Oregon Institute of Marine
Biology Contribution Number 91-002.
Literature Cited
Alspach, G. S. 1972. Osmotic and ionic regulation in the Dungeness
crab. Cancer magister Dana. Ph.D. thesis, Oregon State University,
Corvallis. Oregon.
Charmantier, G., M. Charmantier-Daures, and D. E. Aikens.
1984a. Neuroendocnne control of hydromineral regulation in the
American lobster Homarus amencanus H. Milne-Edwards 1837
(Crustacea, Decapoda) 1. Juveniles. Gen. Comp. Endocrinol. 54:
8-19.
Charmantier, G., M. Charmantier-Daures, and D. E. Aikens.
1984b. Neuroendocnne control of hydromineral regulation in the
American lobster Homarus americanus H. Milne-Edwards 1837
(Crustacea. Decapoda) 2. Larval and postlarval stages. Gen. Comp
Endocrinol. 54: 20-34.
Charmantier, G., P. Thuet, and M. Charmantier-Daures. 1984c. La
regulation osmotique et lontque chez Homarus gammarus (L.)
(Crustacea: Decapoda). / Exp. Mar Biol. Ecol. 76: 191-199.
Charmantier, G., M. Charmantier-Daures, N. Bouaricha, P. Thuet,
D. E. Aiken, and J.-P. Trilles. 1988. Ontogeny of osmoregulation
and salinity tolerance in two decapod crustaceans: Homarus amer-
icanus and Penaeus japonicus Biol Bull. 175: 102-110.
Charmantier, G., and M. Charmantier-Daures. 1991. Ontogeny of os-
moregulation and salinity tolerance in Cancer irroralus: elements of
comparison with C. borealts (Crustacea. Decapoda). Biol. Bull 180:
125-134.
Conte, F. P. 1984. Structure and function of the crustacean larval salt
gland. Int. Rev. Cylol. 91: 45-106.
Engelhardt, F. R., and P. A. Dehnel. 1973. Ionic regulation in the
Pacific edible crab. Cancer magister (Dana). Can. J. Zoo/. 51:
735-743.
Felder, D. L. 1978. Osmotic and ionic regulation in several western
Atlantic Callianassidae (Crustacea, Decapoda, Thalassinidae). Biol.
Bull 154: 409-429.
CRUSTACEAN ION REGULATION ONTOGENY
277
Fclder, J. M., D. L. Felder, and S. C. Hand. 1986. Ontogeny of os-
moregulation in the estuarine ghost shrimp Callianassa jamaicense
var. louisianensis Schmitt (Decapoda, Thalassinidea). J. Exp. Mar.
Biol. Ecol. 99:91-105.
Foskett, J. K. 1977. Osmoregulation in the larvae and adults of the
grapsid crab Sesarma reticulatum Say. Biol. Bull. 153: 505-526.
Gross, W. J. 1964. Trends in water and salt regulation among aquatic
and amphibious crabs. Biol. Bull 127: 447-466.
I lolliday, C. \V. 1980. Magnesium transport by the urinary bladder of
the crab. Cancer magisler. J. Exp Biol. 85: 187-201.
Hunter, K. C., and P. P. Rudy, Jr. 1975. Osmotic and ionic regulation
in the Dungeness crab. Cancer magisler Dana. Camp. Biochem.
Physiol. 51A: 439-447.
Jones, L. L. 1941. Osmotic regulation in several crabs of the Pacific
coast of North America. J. Cell. Comp. Physiol. 18: 79-91.
Kalber, F. A. 1970. Osmoregulation in decapod larvae as a consider-
ation in culture techniques. Helgol. Wiss. Meeresiinlers. 20: 697-
706.
Larimer, J. L., and A. F. Riggs. 1964. Properties of hemocyanins. I.
The effects of calcium ions on the oxygen equilibrium of crayfish
hemocyanin. Comp. Biochem Physiol. 13: 35-46.
Lough, R. G. 1976. Larval dynamics of the Dungeness crab. Cancer
magister. off the central Oregon coast, 1970-1971. Fish. Bull. 74:
353-375.
Mantel, L. H., and L. L. Farmer. 1983. Osmotic and ionic regulation.
Pp. 53-161 in The Biology oj Crustacea, vol. 5, Internal Anatomy
and Physiological Regulation. L. H. Mantel, ed. Academic Press,
New York, London.
Miller, K. I., and K. E. Van Holde. 1974. Oxygen binding in Callianassa
cali/orniensis hemocyanin. Biochemistry 13: 1668-1674.
Rabalais, N. N., and J. N. Cameron. 1985. The effects of factors im-
portant in semi-arid environments on the early development of Uca
subcylindrica. Biol. Bull. 168: 147-160.
Read, G. H. L. 1984. Intraspecific variation in the osmoregulatory ca-
pacity of larval, post larval, juvenile and adult Macrobrachium petersi
(Hilgendorf). Comp. Biochem. Physiol. 78A: 501-506.
Reed, P. H. 1969. Culture methods and effects of temperature and
salinity on survival and growth of Dungeness crab (Cancer magister)
larvae in the laboratory. J. Fish. Res Board Can. 26: 389-397.
Robertson, J. D. 1960. Osmotic and ionic regulation. Pp. 317-339 in
Physiology of Crustacea, (•'<>/. /. T. H. Waterman, ed. Academic Press.
New York.
Sky-Peck, H. H. 1964. Determination of magnesium in serum and
urine. Clin. Chan. 10: 391-398.
Thuet, P., M. Charmantier-Daures, and G. Charmantier. 1988. Relation
entre Osmoregulation et activites d'ATPase Na+-K+ et d'anhydrase
carbonique chez larves et postlarves de Homarus gammarus (L.)
(Crustacea: Decapoda). J. Exp. Mar. Biol. Ecol. 115: 249-261.
Truchot, J.-P. 1975. Factors controlling the in vitro and in vivo oxygen
affinity of the hemocyanin in the crab Carcinus maenas (L.). Respir.
Physiol. 24: 173-189.
Waite, F. C. 1899. The structure and development of the antennal
glands in Homarus americanus Milne-Edwards. Bull. Mus. Comp.
Zool. Han: 35: 151-210 and plates.
Young, A. M. 1979. Osmoregulation in larvae of the striped hermit
crab Clibanarius villains (Box) (Decapoda: Anomura; Diogenidae).
Estuarine Coastal Mar. Set. 9: 595-601.
Reference: Biol. Bull 182: 278-287. (April, 1992)
Visual Rhythms in Stomatopod Crustaceans Observed
in the Pseudopupil
THOMAS W. CRONIN
Department of Biological Sciences, University of Maryland Baltimore County,
Baltimore. Maryland 21228
Abstract. Many aspects of visual function in animals
are influenced by the operation of endogenous rhythms.
Using techniques of intracellular optical physiology, I in-
vestigated visual rhythms in two species of stomatopod
crustaceans (mantis shrimps): Sqiiilla empusa. a species
active throughout the day and night, and Gonodactylm
oerstedii, which is strictly diurnal. Reflectance from within
the deep pseudopupil of the compound eyes and its change
upon stimulation with light were monitored in individual
animals in constant conditions for up to two weeks. Both
species expressed circadian rhythms in visual function.
In S. empusa, the pupillary response was much stronger
during subjective night; little or no response could be elic-
ited during subjective day. In this species, an endogenous
rhythm caused pupillary reflectance to increase during
subjective day. Rhythms in G. oerstedii were of lower
amplitude than in S. empusa and were more difficult to
detect. The differences between these species, together with
the results of other comparative research on visual
rhythms in arthropods, suggest that circadian, rhythmic
processes are involved in optimizing nocturnal eyes for
maximum sensitivity and dynamic range.
Introduction
Endogenous rhythms in visual function are common
among animals. Diverse rhythmic phenomena associated
with vision occur in both vertebrate and invertebrate spe-
cies. For example, in the vertebrates, events known to be
under endogenous control include photoreceptor mem-
brane shedding (LaVail, 1976), retinomotor movements
(reviewed in Levinson and Burnside, 1981; Burnside and
Nagle, 1983), and synthesis of mRNA coding for opsin
(Korenbrot and Fernald, 1989i Invertebrate visual sys-
Received 24 October 1991; accepted 13 January 1992.
terns also express rhythmicity in membrane shedding
(Nassel and Waterman, 1979; Horridge et al. 1981; Wil-
liams, 1982) or preparation for shedding (Chamberlain
and Barlow, 1984). Visual rhythms apparently unique to
invertebrates include cyclic changes in ERG amplitude
(e.g., Arechiga and Wiersma, 1969; Page and Larimer,
1975; Barlow, 1983;FleissnerandFleissner, 1985), in rates
of action potential production (Jacklet, 1969), and, par-
ticularly in arthropods, migration of screening pigment
in secondary pigment cells (Welsh, 1930; Kleinholz, 1937;
Page and Larimer, 1975; see reviews of Stavenga, 1979,
and Autrum, 1981).
In many cases, circadian changes in sensitivity are due
primarily to variations in the quantum catch by the pho-
toreceptor cells, produced by alterations either in asso-
ciated structures such as secondary pigment cells or in
the amount of photoreceptor membrane per cell. But in
some species, the actual photoreceptor cells can undergo
rhythmic changes that affect their ability to respond to
the capture of a photon by rhodopsin. For example, in
Limulus polyphemus. circadian events alter electrophys-
iological properties of individual photoreceptor cells
(Kaplan and Barlow, 1980; Barlow et al., 1987; Kass and
Renninger, 1988).
The photoreceptor cells of many arthropod species are
independently capable of adjusting their sensitivity to light
by mobilizing granules of primary pigment, producing a
phenomenon known as the pupillary response (Kirschfeld
and Franceschini, 1969; see review of Stavenga, 1979).
Arthropod pupillary responses may be observed nonin-
vasively by monitoring light reflected from the deep pseu-
dopupil of the compound eye (Stavenga and Kuiper, 1977;
Bernard and Stavenga, 1979; Cronin, 1989); as the pig-
ment granules migrate inwards in response to photic
stimulation, reflectance rises. Circadian changes in pseu-
dopupillary appearance and level of reflectance have been
278
STOMATOPOD VISUAL RHYTHMS
279
observed in many arthropod compound eyes (Stavenga,
1977; see review of Stavenga, 1979). The rhythms are ap-
parently due to circadian events in secondary pigment
cells, under nervous (Page and Larimer, 1975) or neu-
roendocrine (Smith, 1948; Page and Larimer, 1975; Her-
nandez-Falcon et ai, 1987) control. True pupillary re-
sponses, however, are caused by translocations of primary
pigments, within the actual photoreceptor cells, in direct
response to photic stimulation (Stavenga, 1 979). Do these
responses also express circadian rhythms? If so, the eyes
may be optimized for sensitivity and dynamic range at a
particular phase of the diel cycle, under rhythmic control.
In earlier work with the squilloid stomatopod crusta-
cean Squilla empitsa, I found that it was difficult or im-
possible to elicit any changes in reflection from the deep
pseudopupil during the day, whereas nocturnal stimula-
tion produced large, highly repeatable reflectance increases
(Cronin, 1 989). In contrast, the gonodactyloid stomatopod
species Gonodactylus oerstedii and Pseitdosquilla ciliata
expressed pupillary responses no matter when they were
stimulated. In this report, I describe experiments testing
whether there is a rhythmic component to pupillary func-
tion in S. empusa and G. oerstedii. The results suggest
that rhythmic events strongly alter the pupillary responses
of S. empusa and have a weaker influence on those of G.
oerstedii.
Materials and Methods
Adult animals were used in all experiments. Work with
Squilla empusa was carried out at the Duke University
Marine Laboratory in Beaufort, North Carolina. Animals
were collected locally and maintained either in running
seawater tables exposed to indirect, natural daylight
through windows along two sides of the room (ambient
photoperiod experiments) or in small containers contain-
ing natural seawater placed in a chamber with a controlled
light:dark cycle (reversed photoperiod experiments). An-
imals were fed fresh oyster and shrimp meat. Experiments
with Gonodactylus oerstedii took place in Baltimore, using
animals collected in the Florida Keys. These animals were
kept in aquaria filled with artificial seawater in a 12 h:12
h light:dark cycle, and were fed frozen shrimp.
Reflectance from the deep pseudopupil was monitored
using the techniques of intracellular optical physiology,
described in detail in Bernard and Stavenga (1979) and
Cronin (1989). Dorsal surfaces of animals were attached,
using Scutan dental plastic, to a moveable platform which
was then submersed in seawater. During each experiment,
water in the experimental chamber (which contained
about 1200 ml) was changed occasionally, at irregular
times. Animals were not fed during an experiment.
Once mounted, each experimental animal was aligned
so that the pseudopupil of either the dorsal or ventral half
of the eye (see photographs in Cronin, 1986, and Cronin,
1 989) was centered within the field of view of an incident-
light, photometric microscope. The entire apparatus was
housed in a dark box with a black curtain covering the
front, and experiments took place in a room that was
completely blacked out and isolated from external sources
of light. The central region of the pseudopupil under study,
which appeared to glow dully when viewed by eye, was
isolated using an adjustable field diaphragm. Reflectance
from the pseudopupil was monitored as described in
Cronin (1989), using light of wavelengths >720 nm
(Schott RG720 longpass filter, used with Squilla) or >800
nm (Schott RG800 longpass filter, used with Gonodac-
tylus); this source of light illuminated the eye continually
throughout each experiment, and by itself caused no
measurable pupillary response. At 20- or 30-min intervals,
a stimulating exposure automatically was provided, pro-
duced by passing light from a 1 50-W Xenon arc through
a monochromator (Oriel 7250 with 500-nm blazed grat-
ing), counter-rotating 10-cm diameter neutral density
wedges, and a linear polarizing filter. All stimuli were
confined to the ommatidia contributing to the pseudo-
pupil under study, and were at a wavelength of 500 nm
(half bandwidth of 10 nm). They were produced by open-
ing a Uniblitz electromagnetic shutter under the control
of a microcomputer, and lasted for 5 min (Squilla empusa)
or 30 s (Gonodactylus oerstedii). Measurements of light
reflected from the pseudopupil commenced before each
exposure and continued until well afterwards, and data
were stored on the microcomputer's hard disk for later
analysis. The response for each stimulus was defined as
the average reflectance during the final 20% of the stim-
ulation's duration, divided by the average reflectance be-
fore stimulus onset and following a period in the dark
equal to the duration of the stimulus itself (see also Cronin
and King, 1989). Responses are plotted as the percentage
change in reflectance relative to the average dark levels.
In some cases, the sensitivity of the pupillary response
was measured by providing a series of stimulating inten-
sities over a range of 3.5 to 3.8 density units at steps of
0.5 units. Each series was produced under microcomputer
control, by rotating the neutral density wedges to a series
of preprogrammed settings. The intensity of the stimu-
lating source was measured for each experiment using a
calibrated PIN-10DP/SB photodiode (United Detector
Technology) placed at the position of the animal's eye.
Results
Earlier work with Squilla empusa had revealed that the
ability of an animal to express pupillary responses appar-
ently varied with time. Results of an experiment designed
to detect rhythms in responsiveness are shown in Figure
1. The animal, a male of body length (rostrum-telson)
280
T. W. CRONIN
Day
Figure 1 . Light reflectance from the deep pseudopupil (Reflectance;
top panel) and percentage change in reflectance on stimulation (% Re-
sponse; bottom panel) in an adult male individual of Squilla empusa
maintained in constant conditions. Measurements were made at 30-min
intervals for a total of 327 intervals, from 7 to 14 July 1987. Stimuli
were at 500 nm, at a quantal intensity of 2.9 X 10" quanta cm"2 s~'.
The reflectance was measured as described in the text for 1 min before
each 5-min stimulation; average values are plotted on a scale normalized
to the largest value obtained. Percent response was calculated as described
in the text. Vertical lines are drawn at successive midnights. The light
and dark bands on the abscissa represent times of natural sunrise and
sunset (Eastern Daylight Time). The period (T) of each rhythm, in h, is
given in the top right corner of its panel. Each time series was analyzed
using Enright's periodogram technique (Enright, 1965). Penodogram
amplitudes were computed at 0.1 -h intervals for periods from 10 h to
30 h. The value given on the graph is that of the period in this 20-h
range having the greatest amplitude.
85 mm, was placed in constant darkness in the apparatus
at midday on the first day and stimulated each 30 min
for the next 7 days.
The results of the experiment of Figure 1 are typical of
those of most experiments. Rhythmical variations oc-
curred both in the level of the pupillary response and in
the reflectance from the pseudopupil in the absence of
stimulation. During the subjective day, pseudopupillary
reflectance remained high, and little or no measurable
reflectance change occurred in response to the light stim-
ulus— the variations that were observed were due to ap-
parently random fluctuations. However, near the time of
natural sunset, reflectance from the deep pseudopupil di-
minished, and large increases in reflectance occurred upon
stimulation. Within 1 io 2 h, the reflectance rise during
stimulation changed from near zero to greater than 20%.
Concurrently, baseline pseudopupillary reflectance de-
creased by up to 50%. Near the time of subjective dawn,
the change in reflectance during stimulation dropped once
more to near zero, again over a period of 1 to 2 h, while
the baseline reflectance quickly rose to its daytime level.
This rhythmical pattern was typical of that expressed in
animals maintained under the ambient photoperiod. The
period of the rhythm was estimated using Enright's per-
iodogram technique (Enright, 1965); both rhythms had
periods near 24 h (see Fig. 1; the difference between the
two periods is meaningless with time series of this length).
It is conceivable that the observed rhythms in baseline
reflectance and responsiveness were expressions of a single
phenomenon; the pupillary mechanism could rhythmi-
cally assume its fully light-adapted state during the day,
thus increasing reflectance and losing its ability to adapt
further to light. Two types of observations argue against
this interpretation of the data. First, the two rhythms did
not have exactly inverse forms. Baseline reflectance tended
to change less abruptly than did the level of response, and
on some days its changes were not precisely in phase with
the changes in response level. More convincingly, some
animals expressed rhythms in responsiveness with little
or no circadian change in the baseline reflectance. For
example, in the experiment of Figure 2, the baseline re-
100
V)
.n 80
o C
.y~
^ -
^
^—
— 'X.
I = 25.8
I 60
0 £^
C 2 40
osS
< 20
0 -
14
T = 24.5
\2
l
I
Response
O 00 O
k
fj
V,
A
f\
4
2
o-
J
-.TWIT-I
I
1 - AJ
1
7- '
Day
Figure 2. Results obtained from a male Squilla empusa maintained
in constant conditions from 19 to 23 July 1987. Three series of stimuli
of increasing intensities were given, at the times indicated by arrows, for
determination of response-iwrnv-intensity functions (see text). Except
during these series, stimulation was at a quantal intensity of 1.05 X 10"
quanta cnT: s '. Otherwise as in Figure 1.
STOMATOPOD VISUAL RHYTHMS
281
flectance varied only slightly over four days. The only
changes were a small increase during first dark phase and
transient changes on each successive dusk and dawn.
These results strongly suggest that the ability of the pu-
pillary mechanism itself to respond to light varies rhyth-
mically. Changes in baseline reflectance could not result
from a rhythmical responsiveness to the constant 720-nm
light used to monitor reflectance, for such a rhythm should
produce increasing pseudopupillary reflectance at night,
when responsiveness is greatest. I therefore conclude that
although the two rhythms observed in Figure 1 may be
intimately linked, they are expressions of separate events
within the ommatidia. Once again, periodogram analysis
suggests that the rhythms of Figure 2 are circadian. Per-
iodogram analysis is particularly unreliable for time series
as short as that of Figure 2, but inspection of the forms
of the rhythms clearly confirms that their periods are near
24 h.
The onset of responsiveness at dusk occurs by a smooth
transition (Fig. 3). Although successive responses increase
in size, the time each takes to reach its plateau phase is
similar. It is not primarily the rate of the response that
alters, therefore, but the actual amplitude of the reflectance
change. The increasing sizes of the responses during the
dusk transition mimic the increasing responses that occur
with increasing intensity, when the eye is maximally sen-
sitive (Cronin, 1989). The dusk transition, therefore, ap-
pears to be a time of rapidly increasing sensitivity of the
pupillary mechanism.
The rhythms of responsiveness in these experiments
retained phases that appeared to be tightly linked to the
external diel cycle, and there remained the possibility that
external timing cues were reaching the experimental an-
imals. This possibility was tested by placing an animal in
an isolated chamber subjected to a light:dark cycle phased
differently from the ambient cycle of sunrise and sunset;
lights were switched on at 1800 (Eastern Daylight Time)
and off at 0600 each day. Following 12 days of exposure
to this "inverted" photoperiod, an animal was placed in
constant conditions. Rhythms both of the pupillary re-
sponse and the baseline reflectance were now observed to
be in phase with the imposed cycle. Responsiveness in-
creased, and baseline reflectance decreased, during the
entrained dark period between 6 am and 6 pm. To avoid
repeatedly stressing the animal by mounting it for study,
its visual rhythms were not experimentally defined before
imposing the reversed photoperiod. Nevertheless, this was
the only case in which maximum responses were ever
observed during the astronomical day, so it is reasonable
to conclude that the rhythms had been entrained by the
exposure to the "inverted" photoperiod. Such results also
demonstrate that the rhythms observed in the earlier ex-
periments were not in response to exogenous cues. As
before, the periods of the rhythms were very near 24 h
(Fig. 4).
Response-mros-intensity functions were obtained
from the individuals studied in Figures 2 and 4 during
both the subjective night and the subjective day. The
o
Q.
«
u
Oi
figure 3. Changes in reflectance from the deep pseudopupil of the animal used in the experiment of
Figure 2. during the onset of the nocturnal increase in responsiveness. The 6 traces were obtained at 30-
min intervals, beginning at 1535 EOT (lowest trace) and ending at 1805 (highest trace) on 19 July 1987.
Percent response is computed relative to the average reflectance in the 1 min prior to stimulation. Vertical
lines are drawn at the beginning (0 min) and end (5 min) of the stimulation interval.
282
T. W. CRONIN
100
Day
Figure 4. Results obtained from a female Squilla empusa maintained in constant conditions from I to
1 1 July 1988. Prior to the experiment, the animal was kept in a controlled light:dark cycle as described in
the text; the light and dark bands on the abscissa indicate the times of light and dark of the entraining cycle.
Arrows indicate times of the beginnings of series during which stimulation was increased during each half
hour. At all other times, stimulation was at a quantal intensity of 1.02 X 10" quanta cnT: s~'. Gaps in the
record indicate times of missing data due to equipment failure; otherwise as in Figure 1.
functions were obtained by stimulating the eye on 8 or (|
successive 30-min intervals with stimuli increasing at step;
of 0.5 log units, ultimately providing a maximum quantal
intensity of 3.44 X 10'2 quanta cirT2 s~' (experiment of
Fig. 2) or 2.62 X 1013 quanta cm"2 s"' (experiment of
Fig. 4).
All nighttime series produced response-vm;«-intensity
functions of similar shape, with maximal reflectance in-
creases near 1 3% (Fig. 5A) or 45% (Fig. 5B). In contrast,
the daytime response level remained near 0, rising at most
to about 3% of the height of the nighttime peak at the
maximum stimulation intensity. It appears, in fact, that
the maximum pupillary response that can be generated
is greatly reduced during the day. To test this conjecture
would require stimuli far more intense than what was
available in these experiments. The highest intensities to
which the animal was exposed during these series were
up to 100 times the usual test exposure; somewhat sur-
prisingly, these produced no obvious phase shifts in sub-
sequent cycles of the rhythms (Figs. 2 and 4).
In the earlier work with gonodactyloid stomatopod
species (Cronin, 1989; Cronin and King, 1989), no par-
ticular diel variation in the pupillary responses was no-
ticed. Nevertheless, it seemed likely that gonodactyloids
could possess rhythms similar to those of Squilla, but per-
haps more subtle in their form. This possibility was tested
with Gonodactylus oerstedii, using an overall experimental
design like that employed with Squilla empusa, except
that in these cases entrainment was imposed entirely by
artificial cycles of light and dark. The pupillary response
is much more rapid in G. oerstedii than in S. empusa, so
I used briefer, more frequent stimuli in these experiments.
Results of two experiments, representing the range of ob-
served experimental outcomes, are displayed in Figures
6 and 7.
In the experiment of Figure 6, reflectance from the
pseudopupil prior to stimulation showed a clear circadian
rhythm. Here, the form was rather different from what
was obtained using S. empusa. Soon after the expected
time of lights-out, reflectance from the deep pseudopupil
slowly increased, ultimately becoming about 10% greater
than during the subjective day. Beginning near midday
in the entrainment cycle, this elevated reflectance once
more declined. A rhythm in pupillary responsiveness was
STOMATOPOD VISUAL RHYTHMS
283
12 •
10
o
D.
-3
-2
Log Intensity
Figure 5. Response-vtTOu-intensity functions obtained during the experiments of Figure 2 (A) and
Figure 4 (B). Functions were obtained at the times indicated by arrows on those figures; large circles indicate
results obtained during subjective night and small circles, during subjective day. Open circles correspond to
the first series and closed circles to the second. Photic intensities are relative to 3.44 x 1012 quanta cm"2 s~'
more difficult to detect, although on average, responses
did appear to be slightly larger at subjective night. Per-
iodogram analysis of the data suggests the presence of
circadian cycles in both the baseline reflectance data and
the responsiveness data (Fig. 6).
The experiment of Figure 7 revealed the strongest
expression of circadian rhythms in responsiveness that I
observed in over 20 experiments with G. om/tW//. Am-
plitudes of the pupillary response were about twice as large
during subjective day as during subjective night, thus
having the opposite form of those expressed by 5. empma.
The pupillary response was present both during the day
and night. In this case, however, there was no evidence
of a circadian cycle in pupillary reflectance.
Discussion
As demonstrated here, rhythms with a circadian period
can readily be observed through observations of pseu-
dopupils in compound eyes of stomatopod crustaceans.
In the squilloid species, Squilla empusa. two parallel
rhythms of high amplitude are usually observed: a cycle
in baseline reflectance from the pseudopupil in the absence
of stimulation, and a cycle in the amplitude of the pu-
pillary response to light stimulation. In some experiments,
the gonodactyloid Gonodactylus oerstedii expressed the
baseline reflectance rhythm, although at lower amplitude
than 5. empusa. Rhythms in responsiveness in G. oerstedii
were more difficult to detect than in S. empusa, but they
could be observed in some individuals.
The rhythms that were observed were robust. They
persisted and maintained their form, frequently with a
high amplitude, for a week or more in constant conditions.
The rhythms were clearly circadian; their periods were
consistently revealed to be near 24 h either by inspection
or by periodogram analysis of the data. The rhythm in
responsiveness expressed by S. empusa was particularly
impressive for its rapid rise each subjective evening and
its equally rapid fall in the morning.
During the day, the sensitivity of the pupillary response
decreased by at least three orders of magnitude. These
changes are greater than observed in Limulus polyphemus.
in which both the electroretinogram (ERG) and single-
cell sensitivities decrease by about 1.5 log units (Barlow
et al., 1977, 1987). Other arthropods, however, have sen-
sitivity changes between the day and night states about
as large as those of S. empusa. In the crayfish Cherax
destructor, dark-adapted single photoreceptor cells are
more than two log units more sensitive at night (Bryceson,
1986), while the simple eyes of scorpions rhythmically
gain nearly four log units of sensitivity each night (review:
Fleissner and Fleissner, 1985). In its natural habitat, 5.
empusa probably lacks a pupillary response during the
day. The maximum intensities of stimulation used in
producing the response- irrws-intensity functions of Fig-
284
T. W. CRONIN
_ 100
Figure 6. Results obtained from a female Gonodactylus oerstedii
maintained in constant conditions from 22 to 28 January 1988. The
animal was kept in a controlled light:dark cycle, indicated by the light
and dark bands on the abscissa, prior to the experiment. Stimulation
was at 500 nm, at intervals of 20 min, and at a quantal intensity of 1.57
X 1012 quanta cm'2 s~'. Otherwise as in Figure 1.
ure 4, on the order of 10" quanta cm 2 s ', are similar
to intensities an animal would experience at a depth of
only a few meters in the coastal waters it inhabits (Forward
el al.. 1988). During the daylight phase of the rhythm,
such intensities produced no apparent response.
If these circadian rhythms are to be properly phased to
the diel cycle, they must be entrainable by cycles of light
and dark. Animals apparently do entrain completely to
a novel light:dark cycle within 12 days, as suggested by
the results of the experiment of Figure 4. Presumably,
photoreceptors for this entrainment are either within the
compound eyes or exist elsewhere in the animal. In fact,
many invertebrate species entrain their circadian rhythms
using extraocular pathways (see review of Bennett, 1979).
In crayfish, and probably other decapod crustaceans,
photic entrainment of circadian rhythms can be achieved
by retinal illumination (Larimer and Smith, 1980), but
such entrainment may also involve photoreceptors of the
6th abdominal ganglion (Fuentes-Pardo and Inclan-Ru-
bio, 1987) or other regions of the CNS (Page and Larimer,
1976; Larimer and Smith, 1980). In particular, the work
of Page and Larimer (1976) demonstrated that the caudal
photoreceptors (in the 6th abdominal ganglion) are not
required for entrainment. In contrast to decapod crusta-
ceans, S. empusa lacks this caudal photoreceptor ( Wilkens
and Larimer, 1976), and no other extraretinal photore-
ceptors have yet been described in stomatopods. If en-
trainment is mediated solely by the compound eyes, the
spatially limited stimuli of these experiments (which were
restricted to the ommatidia of the pseudopupil of one
part of a single eye) must have been insufficient to induce
observable phase changes. Nevertheless, intense stimuli
like those used to measure the response-vmzw-intensity
functions produce no obvious phase shifts, implying that
there may be a role for extraocular photoreception in
rhythm entrainment in stomatopods.
What underlying events take place within the com-
pound eyes to bring about the rhythms observed in this
study? Changes in the level of baseline reflectance can be
effected in several ways (Stavenga, 1979). External to the
receptor cells, these include reorganization of optical
structures or associated pigment cells, movement of pig-
ment in secondary pigment cells, and masking or un-
masking of a tapetum. Within the photoreceptors, changes
in reflectance could be caused by alterations in rhabdom
size or microvillar organization, or by events within the
100
Figure 7. Results obtained from a male Gonodactylus oerstedii
maintained in constant conditions from 31 January to 12 February 1991.
The animal was kept in a controlled lightidark cycle, indicated by the
light and dark bands on the abscissa, prior to the experiment. Data from
the first two days of the experiment are not plotted, due to repeated
animal movement and equipment failure. Stimulation was at 500 nm.
at intervals of 20 min, and at a quantal intensity of 1.97 x 10|: quanta
cm~2 s"1. Periodogram analysis of the data of the upper panel (pupillary
reflectance) produced a slowly rising penodogram with a spike peak at
27.1 h; because the analysis produced no clear maximum, the value ot
r is not given on the figure. Otherwise as in Figure 1 .
STOMATOPOD VISUAL RHYTHMS
285
cytoplasm of retinular cells, such as rearrangement of the
perirhabdomal palisade or movements of pigment gran-
ules.
Ommatidial reorganization no doubt accounts for
much of the circadian change in the appearance of the
pseudopupil of Limulus polyphemus (Slavenga, 1979), but
more recent work suggests that the retinular cell pigment
itself migrates under circadian control (Kier and Cham-
berlain. 1990). Long-term studies of anatomical changes
in stomatopod eyes do not exist. In work with the Med-
iterranean species Squilla mantis, SchifF( 1974) stated that
during dark adaptation, rhabdoms increase in length and
decrease in diameter; simultaneously, the crystalline cone
contracts. Schonenberger (1977) also observed the con-
traction of the crystalline cone during dark adaptation,
and noted that the distal pigment cells reorganize at that
time. Within the photoreceptor cells, granules of primary
pigment line up in neat rows, encircling the rhabdom as
it dark adapts. These changes occur during diurnal dark
adaptation, following prolonged exposure to daylight, at
times when Squilla empusa reveals no light-sensitive pu-
pillary responses and no evidence of dark adaptation fol-
lowing stimulation. At present, it is uncertain whether
these two species of Squilla differ in their ability to respond
to light during the day, whether the two sets of studies
involve different phenomena, or whether prolonged ex-
posure to light during the day can in fact produce revers-
ible light-adaptation via the pupillary mechanism.
While cyclic restructuring of ommatidia or transloca-
tion of secondary pigment could produce rhythmic
changes in baseline reflectance, events effecting major al-
terations in the level of the pupillary response must occur
within the actual photoreceptor cells. In apposition eyes
like those of stomatopods, the pupillary response is pro-
duced by radial translocation of granules of primary pig-
ment residing in the photoreceptors (Stavenga. 1979; King
and Cronin, 1989). Reflectance changes may be produced
in superposition compound eyes, however, by events in
secondary pigment cells (Bernard el al., 1984; Weyrauther,
1986). Compared to the responses in stomatopod eyes,
these changes are slow and have considerable inertia —
the process continues, or remains saturated, long after
stimulation ceases.
Migration of primary pigment is directly under the
control of the retinular cell, and, at least in crustaceans,
is thought not to be influenced by hormones (Ludolph et
al., 1973). In another arthropod, Limulus polyphemus,
rhythmic neural input does influence the position of pri-
mary pigment (Kier and Chamberlain, 1990); similar
processes may act in crustacean eyes. The pupillary re-
sponse requires the presence of calcium ions (Kirschfeld
and Vogt, 1980; Frixione and Arechiga, 1981; Howard,
1984). Since excitation of arthropod photoreceptor cells
is also dependent upon intracellular increases in calcium
concentration (see review of Fein and Payne, 1989), the
absence of the pupillary response during the day could
imply that electrophysiological responses of the photo-
receptors are also abolished at that time. However, this is
unlikely to be the case; Kirschfeld and Vogt ( 1 980) showed
that in fly photoreceptors, it is possible to block pigment
migration without changing retinal electrical responses.
In work with mutant flies, Lo and Pak (1981) also ob-
served that electrophysiological responses could remain
in the absence of pigment migration. The processes of
pigment translocation and membrane depolarization,
while both calcium-dependent, are therefore not com-
pletely parallel. The diurnal loss of pupillary responsive-
ness in Squilla empusa could represent another example
of the decoupling of these two processes. Experiments are
desirable in which the electrical and pupillary responses
are monitored simultaneously in this species.
Despite having radically different anatomies, the com-
pound eyes of S. empusa and Limulus polyphemus are,
to some extent, analogous in the functioning of their pu-
pillary responses. Both express rhythmic changes in re-
flectance from the deep pseudopupil, and both show
changes in this reflectance only at night (see Stavenga,
1979). Interestingly, demands on their visual systems
throughout the course of each day may also be analogous.
Neither species is entirely inactive during the day, but
mating of Limulus, a behavior involving vision (Barlow
et al., 1982), occurs mostly during twilight or night (Bar-
low et al., 1986). Indeed, nocturnal vision in Limulus is
extraordinary, enabling the detection of high-contrast tar-
gets under starlight (Barlow et al., 1982). Activity cycles
of Squilla empusa have yet to be examined in the field,
but in the laboratory, at least, this species is most active
during the night (pers. obs.), and the congener, Squilla
mantis, is clearly nocturnal (Froglia and Giannini, 1989).
The ocular design of 5. empusa is that of a nocturnal
animal (Cronin, 1986). It is plausible that the rhythms in
visual function described in this paper are characteristic
physiological properties of a nocturnal compound eye. In
fact, throughout the arthropods, regardless of eye type, all
high-amplitude rhythms in visual physiological function
yet described are in nocturnal species. Besides Limulus,
these include scorpions (simple eyes, rhythm in ERG am-
plitude and sensitivity: reviewed in Fleissner and Fleissner,
1985), crayfish (superposition compound eye, rhythm in
ERG amplitude and sensitivity: Arechiga and Wiersma,
1969; Page and Larimer, 1975; Bryceson, 1986), and
cockroach (apposition compound eye, rhythm in ERG
amplitude: Wills et al., 1986).
G. oerstedii, unlike S. empusa. is active only from dawn
to dusk (Dominguez and Reaka, 1988). At evening twi-
light, it seals off the entrance to its burrow and remains
enclosed all night. Its eyes are optimized for photopic
function; indeed, the specialized spectral receptor classes
286
T. W. CRONIN
in its eyes require reasonably bright light to operate at all
(Cronin and Marshall, 1989; see also Marshall et ai,
1991). Because the eye is used primarily for vision when
photon capture is not limiting, strong rhythmic cycles in
visual function may be unnecessary.
The expression of high-amplitude circadian visual
rhythms in a primarily nocturnal species like S. empusa,
and their absence in the diurnal G. oerstedii, could there-
fore reveal fundamental differences in function between
compound eyes designed for nocturnal (or continuous)
versus diurnal function. If so, the rhythms observed in
this study by monitoring light reflectance from deep pseu-
dopupils are a manifestation of more pervasive underlying
alterations that must occur to maintain visual function
at day and night. The study of rhythmic cycles of sensory
function, and in particular their underlying significance
and control, deserves more attention.
Acknowledgments
This material is based on work supported by the Na-
tional Science Foundation under Grants No. BNS-
85 1 8769 and BNS-89 17183.1 thank D. G. Stavenga and
L. Kass for their comments on an earlier version of the
manuscript.
Literature Cited
Arechiga, H., and C. A. G. Wiersma. 1969. Circadian rhythm of re-
sponsiveness in crayfish visual units. J. Nenrobiol. 1: 71-85.
Autrum, H. 1981. Light and dark adaptation in invertebrates. Pp. 1-
91 in Handbook of Sensory Physiology, Vol VII/6C. H. Autrum, ed.
Springer, Berlin, Heidelberg, New York.
Barlow, R. B., Jr. 1983. Circadian rhythms in the Limulus visual sys-
tem. / Neurosci. 3: 856-870.
Barlow, R. B., Jr., S. J. Bolanowski, Jr., and M. L. Brachman.
1977. Efferent optic nerve fibers mediate circadian rhythms in the
Limulus eye. Science 197: 86-89.
Barlow, R. B., Jr., L. C. Ireland, and L. Kass. 1982. Vision has a role
in Limulus mating behavior. Nature 296: 65-66.
Barlow, R. B., Jr., E. Kaplan, G. H. Renninger, and T. Saito.
1987. Circadian rhythms in Limulus photoreceptors. I. Intracellular
studies. / Gen. Physiol 89: 353-378.
Barlow, R. B., Jr., M. K. Powers, H. Howard, and L. Kass.
1986. Migration of Limulus for mating: relation to lunar phase,
tide height, and sunlight. Biol. Bull 171: 310-329.
Bennett, M. F. 1979. Extraocular light receptors and circadian rhythms.
Pp. 641-663 in Handbook of Sensory Physiology. Vol. VH/6A. H.
Autrum, ed. Springer, Berlin, Heidelberg, New York.
Bernard. G. D., E. D. Owens, and A. V. Hurley. 1984. Intracellular
optical physiology of the eye of the moth Amyelois, J Exp Biol.
138: 155-179.
Bernard, G. D., and D. G. Stavenga. 1979. Spectral sensitivities of re-
tinular cells measured in intact, living flies by an optical method. /
Comp. Physiol 134: 95-107.
Bryceson, K. P. 1986. Diumal changes in photoreceptor sensitivity in
a reflecting superposition eye. / Comp. Physiol. A 158: 573-582.
Burnside, B., and B. Nagle. 1983. Retinomotor movements of pho-
toreceptors and retinal pigment epithelium: mechanisms and regu-
lation. Pp. 67-109 in Progress in Retinal Research, Vol 2, N. Osborne
and G. Chader, eds. Pergamon, Oxford, New York.
Chamberlain, S. C., and R. B. Barlow, Jr. 1984. Transient membrane
shedding in Limulus photoreceptors: control mechanisms under nat-
ural lighting. J. Neurosci. 4: 2792-2810.
Cronin, T. W. 1986. Optical design and evolutionary adaptation in
crustacean compound eyes. J. Crust. Biol. 6: 1-23.
Cronin, T. W. 1989. Application of intracellular optical techniques to
the study of stomatopod crustacean vision. J. Comp. Physiol. A 164:
737-750.
Cronin, T. W., and C. A. King. 1989. Spectral sensitivity of vision in
the mantis shrimp, Gonodactylus oerstedii, determined using non-
invasive optical techniques. Biol Bull 176: 308-316.
Cronin, T. W., and N. J. Marshall. 1989. Multiple spectral classes of
photoreceptors in retinas of gonodactyloid stomatopod crustaceans.
J. Comp. Physiol A 166: 261-275.
Dominguez, J. H., and M. Reaka. 1989. Temporal activity patterns in
reef-dwelling stomatopods: a test of alternative hypotheses. / Exp.
Mar. Biol. Ecol. 117: 47-69.
Enright, J. T. 1965. The search for rhythmicity in biological time-
series. J. Theoret. Biol. 8: 426-468.
Fein, A., and R. Payne. 1989. Phototransduction in Limulus ventral
photoreceptors: roles of calcium and inositol trisphosphate. Pp. 171-
185 in Facets of Vision, D. G. Stavenga and R. C. Hardie, eds.
Springer, Berlin, Heidelberg, New York.
Fleissner, G., and G. Fleissner. 1985. Neurobiology of a circadian clock
in the visual system of scorpions. Pp. 351-375 in Neurobiology of
Arachnids, F. G. Barth, ed. Springer, Berlin, Heidelberg. New York.
Forward, R. B., Jr., T. W. Cronin, and J. K. Douglass. 1988. The
visual pigments of crabs. II. Environmental adaptations. / Comp.
Physiol. A 162: 479-490.
Frixione, E., and H. Arechiga. 1981. Ionic dependence of screening
pigment migrations in crayfish retinal photoreceptors. / Comp.
Physiol 144: 35-43.
Froglia, C., and S. Giannini. 1989. Field observations on diel rhythms
in catchability and feeding of Squilla mantis (L.) (Crustacea, Sto-
matopoda) in the Adriatic Sea. Pp. 221-228 in Biology of Stoma-
topods, E. A. Ferrero, ed. Mucchi, Modena.
Fuentes-Pardo, B., and V. Inclan-Rubio. 1987. Caudal photoreceptors
synchronize the circadian rhythms in crayfish. I. Synchronization of
ERG and locomotor circadian rhythms. Comp. Biochem. Physiol.
86A: 523-527.
Hernandez-Falcon, J., E. Moreno-Saenz, J. M. Farias, and B. Fuentes-
Pardo. 1987. Role of the sinus gland in crayfish circadian rhyth-
micity. I. Pseudopupil circadian rhythm. Comp. Biochem. Physiol.
87A: 111-118.
Horridge, G. A., J. Duniec, and L. Marcelja. 1981. A 24-hour cycle
in single locust and mantis photoreceptors. / Exp. Biol 91: 307-
322.
Howard, J. 1984. Calcium enables photoreceptor pigment migration
in a mutant fly. / Exp. Biol. 113: 471-475.
Jacklet, J. W. 1969. Circadian rhythm of optic nerve impulses recorded
in darkness from isolated eye ofAplysia. Science 164: 562-563.
Kaplan, E., and R. B. Barlow, Jr. 1980. Circadian clock in Limulus
brain increases response and decreases noise of retinal photoreceptors.
Nature 286: 393-395.
Kass, L., and G. H. Renninger. 1988. Circadian change in function of
Limulus ventral photoreceptors. I'isual Neurosci. 1: 3-11.
Kier, C. K., and S. C. Chamberlain. 1990. Dual controls for screening
pigment movement in photoreceptors of the Limulus lateral eye:
circadian efferent input and light. Visual Neurosci. 4: 237-255.
King, C. A., and T. W. Cronin. 1989. Ultrastructural evidence for pu-
pillary response in stomatopod photoreceptors: Cytoskeleton and
orientation of pigment granules. Invest. Ophthalmol I 'isual Sci. 30
(suppl):292.
Kirschfeld, K., and N. Franceschini. 1969. Ein Mechanismus zur Steu-
erung des Lichtfiusses in den Rhabdomeren des Komplexauges von
Musca. Kvbernetik 6: 1 3-22.
STOMATOPOD VISUAL RHYTHMS
287
Kirsi hit-Id, K., and K. Vogt. 1980. Calcium ions and pigment migration
in fly photoreceptors. Naturwissenschaften 67: 516-517.
Kleinholz, L. H. 1937. Studies in the pigmentary system of Crustacea.
II. Diurnal movements of the retina] pigments of Bermudan decapods.
Bio/ Bull 72: 176-189.
Korenbrot, J. 1., and R. D. Fernald. 1989. Circadian rhythm and light
regulate opsin mRNA in rod photoreceptors. Nature 337: 454-457.
Larimer, J. L., and J. T. F. Smith. 1980. Circadian rhythm of retinal
sensitivity in crayfish: modulation by the cerebral and optic ganglia.
J. Comp. Physiol 136: 313-326.
La Vail, M. 1976. Rod outer segment disc shedding in relation to cyclic
lighting. Exp. Eye Res 23: 277-280.
Levinson, G., and B. Burnside. 1981. Circadian rhythms in teleost re-
tinomotor movements. Invest. Ophthalmol. Visual Sci. 20: 294-303.
Lo, M. C, and W. L. Pak. 1981. Light-induced pigment granule mi-
gration in the retinular cells ofDrosophila melanogaster. Comparison
of wild type with ERG-defective mutants. J. Gen. Physiol. 77: 155-
175.
Ludolph, C., D. Pagnanelli, and M. I. Mote. 1973. Neural control of
migration of proximal screening pigment by retinular cells of the
swimming crab. Callinectes sapidus. Biol. Bull. 145: 159-170.
Marshall, N. J., M. F. Land, C. A. King, and T. W. Cronin. 1991. The
compound eyes of mantis shrimps (Crustacea, Hoplocarida, Sto-
matopoda). II. Colour pigments in the eyes of stomatopod crustaceans:
polychromatic vision by serial and lateral filtering. Phil. Trans. R.
Soc. Ser B 334: 57-84.
NasseL, D. R., and T. H. Waterman. 1979. Massive diumally modulated
photoreceptor membrane turnover in crab light and dark adaptation.
J. Comp. Physiol. 131: 205-216.
Page, T. L., and J. L. Larimer. 1975. Neural control of circadian
rhythmicity in the crayfish. II. The ERG amplitude rhythm. / Comp.
Physiol. 97: 81-96.
Page, T. L. and J. L. Larimer. 1976. Extraretinal photoreception in
entrainment of crustacean circadian rhythms. Phoiochem. Photobiol.
23:245-251.
Schiff, H. 1974. A discussion of light scattering in the Squilla rhabdom.
Kybernetik 14: 127-134.
Schonenberger, N. 1977. The fine structure of the compound eye of
Squilla mantis (Crustacea, Stomatopoda). Cell Tissue Res. 176: 205-
233.
Smith, R. I. 1948. The role of the sinus glands in retinal pigment mi-
gration in grapsoid crabs. Bio/. Bull. 95: 169-185.
Stavenga, D. G. 1977. Optics of compound eyes and circadian pigment
movements studied by pseudopupil observations in vivo. Biol. Bull.
153: 446.
Stavenga, D. G. 1979. Pseudopupils of compound eyes. Pp. 357-439
in Handbook of Sensory Physiology, Vol. VII/6A, H. Autrum, ed.
Springer. Berlin, Heidelberg, New York.
Stavenga, D. G., and J. W. Kuiper. 1977. Insect pupil mechanisms. I.
On the pigment migration in the retinula cells of hymenoptera. /
Comp. Physiol. 113: 55-72.
Welsh, J. H. 1930. Diurnal rhythm of the distal pigment cells in the
eyes of certain crustaceans. Proc. Natl. Acad. Sci. 16: 386-395.
Weyrauther, E. 1986. Do retinula cells trigger the screening pigment
migration in the eye of the moth Epheslia kuehniella? J. Comp
Physiol. A 159: 55-60.
Wilkens, L. A., and J. L. Larimer. 1976. Photosensitivity in the sixth
abdominal ganglion of decapod crustaceans: a comparative study. J.
Comp. Physiol. 106: 69-75.
Williams, D. S. 1982. Ommatidial structure in relation to turnover of
photoreceptor membrane in the locust. Cell Tissue Res. 225: 595-
617.
Wills, S. A., T. L. Page, and C. S. Colwell. 1986. Circadian rhythms
in the electroretinogram of the cockroach. J. Biol. Rhythms. 1: 25-
37.
100 YEARS EXPLORING LIFE,
1888-1988
The Marine Biological Laboratory
at Woods Hole
Jane Maienschein
Arizona State University
Essayist Lewis Thomas described the MBL as a "National
Biological Laboratory" that brings together each summer a
collection of biologists from across the United States and
abroad. From its founding in 1888 onward, the MBL has served
as a gathering spot for biologists who come to Woods
Hole not only to work with their favorite marine organisms, but
also to converse with each other and exchange ideas in a way
that seldom happens in the more limited confines of
university biology departments. This wonderful book L> a
biography of the first 100 years of the Marine Biological
Laboratory at Woods Hole, a book that anyone who picks
up will find hard to put down.
1989 • 0-86720-120-7 • 192pp. • Paper • $24.95
JONES AND BARTLETT PUBLISHERS
20 PARK PLAZA
BOSTON, MA 02116
800-832-0034, 617-482-3900 (IN MA)
CONTENTS
POETRY
Skinner, Dorothy ML, and John S. Cook
Carroll M. Williams 165
Mellon, DeForest, Jr.
How the axon got its tale 167
DEVELOPMENT AND REPRODUCTION
- -,--.' . *f -
-
Hand, Cadet, and Kevin R. Uhlinger
The culture, sexual and asexiJal reproduction, and •'<
growth of the sea anemone. Nematostella veclensis 169
McEdward, Larry R.
Morphology and development of a unique type of
pelagic larva in the starfish Prrru\tcr leftflatus (Echn
nodermata: Asteroidea) ...... ,f: .. ~.'. 177
V. '
ECOLOGY AND EVOLUTION
Jeffries, William B., Harold K. Voris, and Sombat
Poovachiranon
Age of the mangrove crab Snlla serrata at coloni-
zation by stalked barnacles of the genus Octulasmis 1 88
Kim, Kiho, Walter M. Goldberg, and George T.
Taylor
Architectural and mechanical properties of the black
coral skeleton (Coelenterata: Antipatharia): a com-
parison of two species 195'
Raimondi, Peter T.
Adult plasticity and rapid larval evolution in a re-
cently isolated barnacle population 210
Shapiro, Daniel F.
Intercolony coordination of zooid behavior and a
new class of pore plates in a marine bryozoan ... 221
Van Alstyne, Kathryn L., Chad R. Wylie, Valerie J.
Paul, and Karen Meyer
Antipredator defenses in tropical Pacific soft corals
(Coelenterata: Alcyonacea). I. Sclerites as defenses
against generalist carnivorous fishes 231'
NEUROBIOLOGY AND BEHAVIOR
Diaz-Miranda, Lucy, David A. Price, Michael J.
Greenberg, Terry D. Lee, Karen E. Doble, and Jose
E. Garcia-Arraras
Characterization of two novel neuropeptides from
the. sea cucumber Holothuria glabernma 241
Mark it-. G. O., C. E. Mills, and C. L. Singla
Giant axons and escape swimming in Euplokamis
duiilapae (Ctenophora: Cydippida) 248
Saigusa, Masayuki
Phase shift of a tidal rhythm by light-dark cycles in
the semi-terrestrial crab Se.wrma pictum 257
PHYSIOLOGY
Baker, S. M., and R. Mann
Effects of hypoxia and anoxia on larval settlement,
juvenile growth, and juvenile survival of the oyster
Crassostrea virginica 265
Brown, A. Christine, and Nora B. Terwilliger
Developmental changes in ionic and osmotic regu-
lation in the Dungeness crab, Cancer magister .... 270
Cronin, Thomas W.
Visual rhythms in stomatopod crustaceans observed
in the pseudopupil 278
Volume 182
THE
Number 3
BIOLOGICAL
BULLETIN
Marine Biological Laboratory"'
LIBRARY
2 2 1992
Hole. Mass
JUNE, 1992
Published by the Marine Biological Laboratory
THE
Marine Biological Laboratory
LIBRARY
JUN221992
BIOLOGICAL BULLETIN
Mass,
PUBLISHED BY
THE MARINE BIOLOGICAL LABORATORY
Associate Editors
PETER A. V. ANDERSON. The Whitney Laboratory. LIniversity of Florida
DAVID EPEL. Hopkins Marine Station, Stanford LIniversity
J. MALCOLM SHICK, University of Maine. Orono
Editorial Board
DAPHNE GAIL FAUTIN. University of Kansas K.. RANGA RAO. University of West Florida
WILLIAM F. GILLY. Hopkins Marine Station, STEVEN VOGEL. Duke University-
Stanford University
Editor: MICHAEL J. GREENBERG, The Whitney Laboratory. University of Florida
Managing Editor PAMELA L. CLAPP, Marine Biological Laboratory
JUNE, 1992
Printed and Issued by
LANCASTER PRESS, Inc.
PRINCE &. LEMON STS.
LANCASTER, PA
THE BIOLOGICAL BULLETIN
THE BIOLOGICAL BULLETIN is published six times a year by the Marine Biological Laboratory, MBL
Street, Woods Hole, Massachusetts 02543.
Subscriptions and similar matter should be addressed to Subscription Manager, THE BIOLOGICAL BUL-
LETIN, Marine Biological Laboratory, Woods Hole, Massachusetts 02543. Single numbers, $30.00. Sub-
scription per volume (three issues), $77.50 ($155.00 per year for six issues).
Communications relative to manuscripts should be sent to Michael J. Greenberg, Editor-in-Chief, or
Pamela L. Clapp, Managing Editor, at the Marine Biological Laboratory, Woods Hole, Massachusetts 02543.
Telephone: (508) 548-3705, ext. 428. FAX: 508-540-6902. E-mail: pamcl@hoh.mbl.edu.
POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN. Marine Biological Laboratory.
Woods Hole, MA 02543.
Copyright © 1992. by the Marine Biological Laboratory
Second-class postage paid at Woods Hole, MA, and additional mailing offices.
ISSN 0006-3185
INSTRUCTIONS TO AUTHORS
The Biological Bulletin accepts outstanding original research
reports of general interest to biologists throughout the world.
Papers are usually of intermediate length (10-40 manuscript
pages). A limited number of solicited review papers may be ac-
cepted after formal review. A paper will usually appear within
four months after its acceptance.
Very short, especially topical papers (less than 9 manuscript
pages including tables, figures, and bibliography) will be pub-
lished in a separate section entitled "Research Notes." A Re-
search Note in The Biological Bulletin follows the format of
similar notes in Nature. It should open with a summary para-
graph of 150 to 200 words comprising the introduction and the
conclusions. The rest of the text should continue on without
subheadings, and there should be no more than 30 references.
References should be referred to in the text by number, and
listed in the Literature Cited section in the order that they appear
in the text. Unlike references in Nature, references in the Re-
search Notes section should conform in punctuation and ar-
rangement to the style of recent issues of The Biological Bulletin.
Materials and Methods should be incorporated into appropriate
figure legends. See the article by Lohmann et at. (October 1990.
Vol. 179: 2 14-2 1 8) for sample style. A Research Note will usually
appear within two months after its acceptance.
The Editorial Board requests that regular manuscripts con-
form to the requirements set below; those manuscripts that do
not conform will be returned to authors for correction before
review.
1. Manuscripts. Manuscripts, including figures, should
be submitted in triplicate. (Xerox copies ol photographs are not
acceptable for review purposes.) The original manuscript must
be typed in no smaller than 12 pitch, using double spacing (///-
eluding figure leg ' -K footnotes, bibliography, etc.) on one side
of 16- or 20-lb. bon • paper, 81/: by 1 1 inches. Please, no right
justification. Manusi. ;; should be proofread carefully and er-
rors corrected legibly • i I i:ick ink. Pages should be numbered
consecutively. Margins on all sides should be at least 1 inch (2.5
cm). Manuscripts should conform to the Council oj Biology Ed-
itors Style Manual. 5th Edition (Council of Biology Editors.
1983) and to American spelling. Unusual abbreviations should
be kept to a minimum and should be spelled out on first reference
as well as defined in a footnote on the title page. Manuscripts
should be divided into the following components: Title page.
Abstract (of no more than 200 words). Introduction, Materials
and Methods. Results, Discussion, Acknowledgments, Literature
Cited. Tables, and Figure Legends. In addition, authors should
supply a list of words and phrases under which the article should
be indexed.
2. Title page. The title page consists of: a condensed title
or running head of no more than 35 letters and spaces, the
manuscript title, authors' names and appropriate addresses, and
footnotes listing present addresses, acknowledgments or contri-
bution numbers, and explanation of unusual abbreviations.
3. Figures. The dimensions of the printed page, 7 by 9
inches, should be kept in mind in preparing figures for publi-
cation. We recommend that figures be about 1 '/: times the linear
dimensions of the final printing desired, and that the ratio of
the largest to the smallest letter or number and of the thickest
to the thinnest line not exceed 1:1.5. Explanatory matter gen-
erally should be included in legends, although axes should always
be identified on the illustration itself. Figures should be prepared
for reproduction as either line cuts or halftones. Figures to be
reproduced as line cuts should be unmounted glossy photo-
graphic reproductions or drawn in black ink on white paper,
good-quality tracing cloth or plastic, or blue-lined coordinate
paper. Those to be reproduced as halftones should be mounted
on board, with both designating numbers or letters and scale
bars affixed directly to the figures. All figures should be numbered
in consecutive order, with no distinction between text and plate
figures. The author's name and an arrow indicating orientation
should appear on the reverse side of all figures.
4. Tables, footnotes, figure legends, etc. Authors should
follow the style in a recent issue of The Biological Bulletin in
preparing table headings, figure legends, and the like. Because
of the high cost of setting tabular material in type, authors are
asked to limit such material as much as possible. Tables, with
their headings and footnotes, should be typed on separate sheets,
numbered with consecutive Roman numerals, and placed alter
the Literature Cited. Figure legends should contain enough in-
formation to make the figure intelligible separate from the text.
Legends should he typed double spaced, with consecutive Arabic
numbers, on a separate sheet at the end of the paper. Footnotes
should be limited to authors' current addresses, acknowledg-
ments or contribution numbers, and explanation of unusual
abbreviations. All such footnotes should appear on the title page.
Footnotes are not normally permitted in the body of the text.
5. Literature cited. In the text, literature should be cited
by the Harvard system, with papers by more than two authors
cited as Jones el til., 1980. Personal communications and ma-
terial in preparation or in press should be cited in the text only,
with author's initials and institutions, unless the material has
been formally accepted and a volume number can be supplied.
The list of references following the text should be headed Lit-
erature Cited, and must be typed double spaced on separate
pages, conforming in punctuation and arrangement to the style
of recent issues of The Biological Bulletin. Citations should in-
clude complete titles and inclusive pagination. Journal abbre-
viations should normally follow those of the U. S. A. Standards
Institute (USASI). as adopted by BIOLOGICAL ABSTRACTS and
CHEMICAL ABSTRACTS, with the minor differences set out below.
The most generally useful list of biological journal titles is that
published each year by BIOLOGICAL ABSTRACTS (BIOSIS List of
Serials; the most recent issue). Foreign authors, and others who
are accustomed to using THE WORLD LIST OF SCIENTIFIC PE-
RIODICALS, may find a booklet published by the Biological
Council of the U.K. (obtainable from the Institute of Biology,
41 Queen's Gate, London, S.W.7, England, U.K.) useful, since
it sets out the WORLD LIST abbreviations for most biological
journals with notes of the USASI abbreviations where these differ.
CHEMICAL ABSTRACTS publishes quarterly supplements of ad-
ditional abbreviations. The following points of reference style
for THE BIOLOGICAL BULLETIN differ from USASI (or modified
WORLD LIST) usage:
A. Journal abbreviations, and book titles, all underlined
(for italics)
B. All components of abbreviations with initial capitals
(not as European usage in WORLD LIST e.g.. J. Cell. Camp.
Physiol. NOT/ cell. camp. Physio/.)
C. All abbreviated components must be followed by a pe-
riod, whole word components musl not (i.e.. J Cancer Res.)
D. Space between all components (e.g.. J. Cell. Camp
Physio!., not .I.Cell.Camp 1'hysial.)
E. Unusual words in journal titles should be spelled out
in full, rather than employing new abbreviations invented by
the author. For example, use Rit I 'isindajjelags Islendinga with-
out abbreviation.
F. All single word journal titles in full (e.g., I'eliger, Ecol-
iigy. Brain).
G. The order of abbreviated components should be the
same as the word order of the complete title (i.e.. Proc. and
Trans, placed where they appear, not transposed as in some
BIOLOGICAL ABSTRACTS listings).
H. A few well-known international journals in their pre-
ferred forms rather than WORLD LIST or USASI usage (e.g..
Nature. Science. Evolution NOT Nature. Land.. Science, N.Y.:
Evolution. Lancaster. Pa.)
6. Reprints, page proofs, and charges. Authors receive their
first 100 reprints (without covers) free of charge. Additional re-
prints may be ordered at time of publication and normally will
be delivered about two to three months after the issue date.
Authors (or delegates for foreign authors) will receive page proofs
of articles shortly before publication. They will be charged the
current cost of printers' time for corrections to these (other than
corrections of printers' or editors' errors). Other than these
charges for authors' alterations. The Biological Bulletin does not
have page charges.
Reference: Biol. Bull 182: 289-297. (June, 1992)
Evidence for a Programmed Circannual Life Cycle
Modulated by Increasing Daylengths in Neanthes
limnicola (Polychaeta: Nereidae) From
Central California
PETER P. FONG1 AND JOHN S. PEARSE
Biology Board of Studies and Institute of Marine Sciences. University of California,
Santa Cruz, California. 95064
Abstract. Timing of parturition, fecundity, and life span
were determined in laboratory cultures of the semelparous,
self-fertilizing, viviparous polychaete Neanthes limnicola.
Worms were exposed to fixed daylengths (short — 8h light:
16hdark; neutral— 12h:12h; long— 16h:8h), switched be-
tween different fixed daylengths, and switched from fixed
daylengths to increasing or decreasing daylengths. Timing
of parturition was synchronized when under neutral day-
length, but became asynchronous under both short and
long daylength, as well as when any of the fixed daylength
was followed by decreasing daylengths. Worms under
neutral daylength had the highest fecundities and shortest
life spans, while those under long days had the lowest
fecundities and longer life spans. When fixed daylength
(short, neutral, long) was followed by increasing day-
lengths, timing of parturition was synchronized, fecundity
was high, and life span shortened. These and earlier pub-
lished experiments on the influence of seasonally changing
photoperiods indicate that the life cycle of the estuarine
A', limnicola is programmed to be completed in somewhat
less than a year, and that seasonally changing photoperiods
modulate it to determine the optimal time of parturition.
Introduction
Photoperiodic control of reproduction has been dem-
onstrated for many organisms (reviewed by Saunders,
1982: Gwinner, 1986). Most experimental work has fo-
cussed on the effects of fixed daylength on annual repro-
Received 17 September 1991; accepted 18 February 1992.
' Present address: Department of Physiology. Wayne State University,
Detroit. MI 48201.
ductive rhythms (e.g., for annelids: Garwood and Olive,
1982; Olive and Filial, 1983; Clark, 1988; Schierwater
and Hauenschild, 1990). However, seasonally changing
photoperiod also has profound effects on the timing of
reproduction (Goss, 1982, 1984; Pearse et a!.. 1986). Such
is the case for Neanthes limnicola (Johnson, 1901 ), a vi-
viparous, self-fertilizing hermaphroditic polychaete that
gives birth mainly during the spring (late February-May)
in the brackish-water creeks flowing into Monterey Bay,
California (Smith, 1950; Fong and Pearse, 1992). This
semelparous worm responds to seasonally changing pho-
toperiod by giving birth to young in spring light regimes
when maintained under either in-phase or 6 months out-
of-phase light conditions. Worms in culture live for 6
months to 2 years and still reproduce mainly in spring
light regimes (Fong and Pearse, 1992). These findings
suggested to us that the worms must "see" either one or
more critical daylengths, or increasing daylengths mim-
icking spring light regimes, to complete sexual maturation.
The present paper reports on experiments that examined
the effects of constant fixed daylength, and fixed day-
lengths followed by increasing or decreasing daylengths,
on the timing of parturition, on fecundity, and on life
span in culture of N. limnicola. These experiments re-
vealed evidence for an endogenous, circannual rhythm
that responds to increasing daylengths.
Materials and Methods
We have maintained worms and their offspring in the
laboratory since October 1987 when approximately 20
adults were collected from Watsonville Slough, California
(36°45'N; 121°45'W). Worms used in all experiments
289
290
P. P. FONG AND J. S. PEARSE
were born in the laboratory as a result of self-fertilizations
from lab-reared adults. Although worms in the field are
usually born in the spring, birth date in the laboratory
can be controlled by manipulating seasonally changing
daylengths(Fongand Pearse, 1992). Juvenile worms were
reared singly, initially in small, plastic petri dishes, then
transferred to 80 X 100 mm pyrex culture dishes with lids
( 1 worm/dish). All culture dishes were maintained at lab-
oratory air temperature (Fig. 1), and a salinity of 15%o.
Worms were fed brine shrimps once per week, and the
culture media was changed 1 -2 days after each feeding.
As worms grew and matured they were monitored daily
for signs of reproduction. About 10 days before giving
birth, the body wall of the adult becomes greenish and
semi-transparent, and developing embryos are easily seen
moving through the coelomic fluid. At birth, juvenile
worms emerge through fissures in the degenerating body
wall of the dying adult. Occasionally, adult worms with
normal body size and reproductive morphology produced
no young. Because they looked and behaved normally
before parturition, these worms (n = 7) were included in
all statistical analyses. Parturition date, life span (days
adults spent in culture) and fecundity (numbers of young
produced) were recorded for each birth. All experiments
(treatments and numbers of worms used) are summarized
in Table I. For statistical analysis of parturition dates,
sequential numerical values were assigned to each date
in each experiment; the value of 1 was assigned to the
date of the first parturition. 2 to the following calendar
day, etc. (e.g.. in experiment A the first parturition oc-
curred on 2 April 1 989 and was given the value 1 . 3 April
= 2, and 5 May = 33. etc.). Analysis of variance was used
to compare mean parturition dates, life span, and fecun-
dity (see Fong, 1991, for full analysis).
Fixed daylength (experiment A)
Immediately after birth (Aug-Sept 1988), each of the
108 worms was placed in one of three light-tight wooden
boxes (n = 36/box) illuminated with fluorescent lights
(General Electric F40 daylight) at fixed photoperiods of
either long (L:D 16h:8h), neutral (12h:12h). or short
(8h:16h) daylengths. The northern distribution of N.
limnicola extends to Vancouver Island, where extreme
daylengths of 16h:8h and 8h:16h occur in June and
December, respectively. In central California, extreme
daylengths are approximately 14.5h:9.5h and 9.5h: 14. 5h.
After 3 months, some of the worms in each light treatment
were removed from their boxes and switched into one of
the other two light regimes (for example, 24 worms were
removed from the neutral daylength box, and 12 each
were placed under short and long daylengths; symbolized
by neutral -»• short and neutral -»• long). The remainder
of the worms were maintained under their initial fixed-
o
Q)
0. 16
0)
ASONDJ FMAMJ J ASONDJ FMAMJ J ASON
1988 1989 1990
Figure 1. Monthly air temperature (x + S.D.) at Long Marine Lab-
oratory from daily records.
daylength light regimes. Of the 108 worms. 27 (25%) died
during the experiments either before or after shifting light
regimes (Table I).
Fixed daylength followed hy increasing daylengths
(experiment B) or by decreasing daylengths
(experiment C)
In experiment B, 60 worms (all born on 4 May 1989)
were divided equally among 3 fixed daylength light re-
gimes of long (L:D 16:8), neutral (12:12), or short (8:16)
daylength. After 3 months, roughly half of the worms in
each light treatment were switched into a room where
daylengths would be increasing for almost 5 months, (light
regimes corresponding to February-June), controlled by
a mechanical warehouse clock switch (Astronomic Time
Switch, R.W. Cramer & Co., Type SY Model SOL). The
lights were turned on and off at local sunrise and sunset 6
months out of phase with ambient photoperiod. The early
February light regime was approximately (L:D) 10:14, thus
worms initially exposed to neutral daylength (12:12) saw
a shorter, but increasing daylength for the first 6 weeks
after shifting. Those initially exposed to long ( 16:8) day-
length also saw shorter, but increasing daylengths. Only
worms initially exposed to short (8:16) daylength saw
longer and increasing daylengths. The other half of the
worms in each treatment were maintained in their original
fixed-daylength light regimes (control). Of the 60 worms.
17 (28rr ) died either before or after shifting (Table I).
In experiment C. 60 worms (born 30 July-5 August
1 989) were kept in phase (decreasing daylength) for about
1 month after birth, then divided equally and maintained
as above in either short, neutral, or long daylengths for 7
weeks. In late September. 10 worms from each treatment
were transferred to a room where daylengths (controlled
by another time switch) would be decreasing (in phase)
A PROGRAMMED LIFE CYCLE IN NEANTHES
Table I
Menu I'urtiintiiin (Idles, menu ntiinhcrs n/ ynitni;. und mean ilnv\ in culture <>/ Neanthes limnicola in nil i'A/>iT//mws
291
Experiment name Photoperiod treatment
Date adults
horn
n
surviving
Mean parturition date
Mean #
young
(+S.D.)
Mean life
span (+S.D.)
A. Fixed daylength short
22 Aug 88
6
4 May
100.6 (60.4)
252.6(18.0)
neutral
22-28 Aug 88
8
12 July
163.8 (44.3)
318.5 (141.8)
long
22 Aug 88
10
1 1 Jill)
114.4 (44.5)
313.3(38.5)
short -» neutral
22 Aug 88
7
3 June
124.4 (41.0)
281.4(92.3)
short -» long
22 Aug 88
7
10 June
124.1 (49.5)
286.4 (59.8)
neutral -» short
28 Aug 88
11
1 1 May
153.7 (25.7)
254.2 (15.1)
neutral — » long
28 Aug.
12
10 Aug
99.91 (56.1)
340.2 (97.1)
1 4 Sept 88
long -» short
22 Aug 88
9
1 1 July
139.7 (55.8)
309.6 (55.8)
long -*• neutral
22 Aug 88
1 1
15 Aug
92.3 (77.4)
349.9(102.6)
Independent of shifting
initially short
20
28 May
117.2 (48.8)
274.4(64.2)
initially neutral
31
1 July
135.5 (51.7)
304.0 (98.6)
initially long
30
24 July
113.9 (60.0)
325.6 (72.8)
Pooled
pooled short + neutral
51
18 June
128.3 (50.9)
292.4(87.3)
pooled long
30
24 July
113.9 (60.0)
325.6(72.8)
B. Fixed -» short
4 May 89
6
1 Apr
87.2 (68.7)
332.6(141.5)
increasing neutral
4 May 89
7
1 Apr
143.4 (66.0)
330.3(18.7)
daylength long
4 May 89
4
23 May
64.6 (56.9)
385.0(113.0)
short -* inc
4 May 89
7
29 Nov ( = May light regime)
105.8 (48.5)
209.4(21.5)
neutral -» inc
4 May 89
10
14 Dec (=June light regime)
115.5 (35.8)
223.3(8.1)
long -* inc
4 May 89
9
12 Jan (=July light regime)
152.8 (46.5)
253.1 (7.4)
C. Fixed -» short
5 Aug 89
9
22 July
89.22 (55.9)
350.7(116.1)
decreasing neutral
5 Aug 89
9
27 April
108.9 (36.7)
265.0(5.1)
daylength long
30 July 89
5
14 Aug
39.4 (41.2)
380.4 (32.7)
short -» dec-
5 Aug 89
10
20 Aug
89.2 (40.4)
377.3(131.8)
neutral — » dec-
5 Aug 89
8
20 May
106.3 (18.5)
288.3 (25.4)
king — » dec-
30 July 89
8
22 June
77.87 (30.2)
327.3(37.6)
Independent of shifting
initially short
19
6 Aug
89.2 (46.9)
364.7 (121.9)
initial!) neutral
17
8 May
107.6 (28.7)
275.9(21.0)
initially long
13
13 Jul\
63.1 (38.4)
347.7 (43.7)
Parturition dates are dates on which adults ga\e birth. Daylengths are: short daylength = (L:D 8h:16h). neutral daylength = (12h:12h),
long daylength = (L:D 16h:8h). Short -» neutral indicates cultures which were initiated in (8:16), then after 3 months shifted into 12:12. etc.
inc = increasing daylengths; dec = decreasing daylengths.
for about 3 months. On 23 December 1989, worms were
placed in constant short daylengths (8:16) where they re-
mained until they gave birth. Control worms (n = 10 for
each light treatment) were maintained in their original
fixed-daylength light regimes. Of the 60 worms, 1 1 (18%)
died during the experiment (Table I).
Results
Fixed daylength (experiment A)
Photoperiod significantly affected the timing of par-
turition in Neanthes limnicola (one-way ANOVA of mean
parturition date, F8 72 = 2.08, P = 0.04). Worms main-
tained in constant short and neutral -*• short daylengths
gave birth on average in the ambient spring (May), and
those in short -»• neutral and short -»• long in the ambient
late spring (June) (Table 1, Fig. 2). Worms maintained in
constant neutral (12:12) also gave birth in the ambient
spring, but parturition dates were separated by 10-11
months (Fig. 2); thus the mean parturition date was in
July. Those worms that encountered long daylengths ini-
tially, gave birth in July and August on average (Table I,
Fig. 2). Four significant comparisons were made be-
tween groups (Fisher's Least Significant Difference test, P
< 0.05), and all involved groups that saw either short or
neutral daylengths initially, compared with those that saw
long daylengths at any stage. In each case, exposure to
long daylength resulted in worms giving birth significantly
292
P. P. FONG AND J. S. PEARSE
1989 1990
MA.MJJASONDJFMAMJJASO
Short
Neutral
Long
Short-->Neutral
Short- ->Long
Neutral -->Short
Neutral-- >Long
Long-->Short
Long-- > Neutral
(n=6)
(n=8)
(n=10)
• • « • • • • (n=7)
(n=7)
(n=12)
(n=9)
Experiment A
Figure 2. Parturtion dates of Niwilhcs limiucolii in various conditions of fixed daylength (experiment
A). Each point represents a parturition. Daylengths are short = (L:D 8:16), neutral = (L:D 12:12), and long
= (L:D 16:8). Short -> neutral indicates cultures initiated in short daylength. then shifted into neutral
daylength. etc. Adults that gave birth were themselves born in August-September 1988.
later in the calendar year, and spread over an extended
period. Despite significant differences in the timing of
parturition, photoperiod did not significantly affect life
span (one-way ANOVA, F8.72 = 1.75 P = 0.10), but did
have a significant effect on fecundity (one-way ANOVA,
F8 72 = 2.20 P = 0.04). Highest fecundities were recorded
in worms that saw either constant neutral (x = 163.8, n
= 8) or neutral -* short (x = 153.7, n = 11) day lengths
(Table I). Lowest fecundity was in the long -»• neutral
treatment (x = 92.27. n = 1 1): 4 of these 1 1 worms had
good reproductive morphology, but produced no young.
The initial daylengths to which worms were exposed
(i.e.. independent of shifting) had a significant effect on
timing of parturition (one-way ANOVA of mean partu-
rition dale F- 78 = 3.03, P = 0.05); worms in short and
neutral daylengths gave birth in the spring and early sum-
mer (xshort = May. \neutrai = 1 July, Table 1), but worms
that initially saw long daylengths gave birth in the mid-
summer (X|0ng = 24 July; Fisher's LSD, P < 0.05 for both
comparisons). However, no significant difference in life
span or fecundity exists among these photoperiodic
groups.
Increasing daylenxt/i experiments (experiment B)
The timing of parturition was affected by increasing
daylength (one-way ANOVA of mean parturition date,
F5 37 = 7.79. P = 0.0001). Worms that saw increasing
daylength gave birth in late spring-early summer light
regimes (xshon^,nc = 29 November = May light regime,
Xneutrai-mc = I4 December = June light regime, x,ong^in<;
= 12 January = July light regime; Table 1). and in each
treatment, most of the parturition dates were clustered
within 1-2 months of each other (Fig. 3). Worms in con-
stant neutral daylength showed a trend similar to those
that saw increasing daylengths, giving birth in the ambient
spring (x = 1 April) with all births clustered within 2
months of each other (Fig. 3). Most worms in constant
short daylength gave birth from October 89 to May 1990
(x = 1 April); one worm lived in culture until November
1990, had good reproductive morphology, but produced
no young. Of the four worms in constant long daylength.
three gave birth from June to September 1990 (x = 22
Aug.); one worm lived until late December 1989 but pro-
duced no young. In the latter two light treatments, no
consistent pattern of parturition timing was evident.
A PROGRAMMED LIFE CYCLE IN NEANTHES
293
1989 1990
S ONDJFMAMJJASONDJF M
Short
Neutral
Long
Short-->inc
Neutral-->inc
Long--> inc
(n=6)
[n-7]
(n=4)
(n=7)
(n=10)
Experiment B
1990
MAM
J J
S 0 N D
1991
J F M
Short
Neutral
Long
Short-->dec
Neutral-->dec
Long --> dec
(n=9)
(n=5)
(n=10)
.: (n=8)
(n=8)
Experiment C
Figure 3. Panurition dates of Meant/lex limnicola exposed to constant short (L:D 8:16), neutral (12:12),
or long (16:8) daylength, initial fixed daylength followed by increasing daylength (upper: expenment B),
and initial fixed daylength followed by decreasing daylength (lower: experiment C). Worms exposed to
increasing daylengths were 6 months out of phase with ambient daylength, thus parturition dates in October,
November, December, and January were actually in light regimes corresponding to April, May, June,
and July.
Three months of fixed daylength followed by increasing
daylength had a significant effect on life span (Table I;
one-way ANOVA, F5,37 = 7.73, P = 0.0001 ). Worms that
saw increasing daylength spent the shortest time in culture,
and all three pair-wise comparisons of fixed (control) ver-
sus increasing daylengths (e.g., fixed short daylength versus
short -*• increasing daylength) showed significant differ-
ences (Fisher's LSD, P < 0.05 for all three comparisons).
Increasing daylength had a significant effect on fecun-
dity (one-way ANOVA, F5 ,7 = 7.79. P = 0.05), yet there
was no consistent trend (Table I). The highest mean
fecundity recorded was in cultures that experienced
long -*• increasing daylengths (x = 152.77), even though
these worms lived from 77 to 132 days less than all three
of the fixed-daylength treatments (Fig. 4). High mean fe-
cundity (x = 143.4) also was recorded in constant neutral
294
P. P. FONG AND J. S. PEARSE
8:16 12:12 16:8 8:16->inc 12:12->rnc 16:8->mc
Photoperiodic regime
Figure 4. Fecundity (number of young horn) and life span (days in
culture) in experimental light regimes of experiment B. inc: increasing
daylengths.
daylength, but the lowest mean fecundities were in con-
stant long (x = 64.75) and constant short (x = 87.16)
daylengths.
Comparison of pooled constant daylength (constant
short, neutral, and long) with pooled increasing daylength
showed a significant difference in life span. Worms that
saw increasing daylengths independent of their initial fixed
daylength, reproduced at a younger age than those in con-
stant daylength (xconstam = 344 days, n == 17, xjncreasmg
= 230 days, n = 26; t = 5.83, df = 4 1 , /> = 0.000 1 ).
Decreasing daylength experiment (experiment C)
Seven weeks of fixed daylength followed by two months
of decreasing daylength, and then constant short daylength
had a significant effect on the timing of parturition (Table
I; one-way ANOVA of mean parturition date, F5 43 = 2.83,
P = 0.02). Cultures in neutral and neutral — »• decreasing
daylength reproduced in the ambient spring and showed
much tighter reproductive synchrony than cultures in
other light treatments, which reproduced later in the year,
on average, and with much greater spread in the partu-
rition dates (Fig. 3).
Life span was affected by decreasing daylengths (Table
I; one-way ANOVA, F543 = 2.76, P = 0.03). Those worms
in cultures maintained at neutral and neutral -»• decreas-
ing daylengths took the shortest time to reproduce (xnculral
= 265 days, xneulra|^decrcasing = 288 days).
Decreasing daylengths affected fecundity (Table I; one-
way ANOVA, F5.43 = 2.50, P = 0.04). Highest fecundities
were recorded in cultures at neutral (x = 108.9 young)
and neutral — » deci easing (x = 106.3 young), even though
on average, worms in both these treatments had a shorter
life span than worms in cultures in other light treatments.
Worms in long daylength had the longest life span, but
produced the fewest young (x = 39.4).
Initial fixed daylength treatments, independent of
shitting, had a significant effect on all three parameters.
Mean parturition dates are significantly different (one-
way ANOVA, F:.46 = 5.99, P = 0.004); worms initially
exposed to neutral daylengths had a mean parturition date
in spring (x = 8 May) whereas those initially exposed to
short daylength gave birth in spring and fall (xshort = 6
August) and those in long daylength reproduced in the
summer and fall (x"iong = 13 July) (Fig. 3; Fisher's LSD,
P < 0.05 for both comparisons). Correspondingly, worms
exposed to neutral daylength initially have shorter life
spans than those exposed initially to either short or long
daylength (xneulra| = 276 days, xshort = 365 days, xlong
= 348 days; one-way ANOVA, F: 46 = 5.92, P = 0.005;
Fisher's LSD, P < 0.05 for both comparisons). Fecundity
was also significantly affected (one-way ANOVA, F2,46
= 4.77; P = 0.01 ); worms initially exposed to neutral day-
length produced significantly more young (x = 107.64)
than worms initially exposed to long (x = 63.07; Fisher's
LSD, P < 0.05). but not to short (x = 89.21) daylengths.
Discussion
We have shown that seasonally changing photoperiod
controls the timing of parturition in Neanthes limnicola
from central California (Fong and Pearse, 1992). In the
field, worms give birth mainly in late winter-spring (late
February-May), and in the laboratory, parturition can be
shifted to late summer-fall when the worms had been
reared under seasonally changing photoperiods 6 months
out of phase with ambient. In California, winter-spring
light regimes increase from about (9.5:14.5 L:D) on 21
December to about (14.5:9.5 L:D)on 21 June. Thus, most
worms experience increasing daylengths for 2-5 months
before giving birth.
In the present study, worms exposed to increasing day-
lengths (corresponding to changes in light regimes from
February to June) after 3 months of either fixed short,
neutral, or long daylengths gave birth within 3-5 months,
independent of the initial fixed daylengths to which they
were exposed. That increasing daylengths act to synchro-
nize parturitions in N. limnicola corresponds to our earlier
findings (Fong and Pearse, 1992).
Parturition also was synchronized when the worms were
exposed to fixed, neutral (12:12) daylength: nearly all gave
birth at 9-11 months of age in all three experiments. In
experiment A. most worms in neutral daylength gave birth
at 8-10 months of age. but two worms gave birth 10-12
months later in the following late winter-spring. The latter
two worms may have missed the "gate-open period,"
which specifies a time interval in which worms may ini-
tiate a rapid phase of oocyte growth (Olive, 1984), and
had to wait another full cycle for it to reopen (see below).
However, these worms never saw any changes in photo-
period with which to gauge time and synchronize repro-
duction. Laboratory temperatures did vary, but with little
260
240
220
200
180
160.
140.
120.
100.
80.
60.
40.
20.
A PROGRAMMED LIFE CYCLE IN XEANTHES
Spearman Rank Correlation, r=-0 248, N.173 P.oooi
295
350 400 450
Days in culture
Figure 5. Regression ot'numher of young on days in culture. Data are combined from all three experiments.
pattern by which the animals could seasonally synchronize
activities (Fig. 1 ). Parturition by most worms under fixed,
neutral daylength was in April-June 1989 in experiment
A, following increasing temperatures, and in April-May
1990 in experiments B and C, following a period of little
temperature change. Previous experiments with Ar. lim-
nicola (Fong and Pearse, 1992) showed that temperature
had no effect on the timing of parturition in worms ex-
posed to seasonally changing daylength from birth.
Constant short (8:16) and long (16:8) daylength had
no consistent effect on parturition synchrony but tended
to desynchronize parturition especially when long day-
length was combined (either before or after) with neutral
daylength. Short and long daylengths also had a detri-
mental effect on fecundity. However, the disruptive effect
of long daylength on reproductive timing and fecundity
was mitigated when followed by increasing daylengths.
Decreasing daylengths did not synchronize parturition
and no consistent trend was evident. Moreover, partu-
rient synchronization was not as tight among worms in
neutral -*• decreasing daylengths as it was in constant
neutral daylengths (Fig. 3).
Exposure to different photoperiodic regimes affected
fecundity. The highest fecundity in any experimental
treatment was found in fixed neutral daylength (experi-
ment A), and the lowest in long daylength (experiment
C). Nevertheless, fecundities in the present experiments,
in general, were lower than those previously recorded in
AT. limnicola exposed to seasonally changing daylengths
continuously from birth (Fong and Pearse, 1992). Thus,
a seasonal cycle of decreasing and increasing daylengths
may be necessary for maximum fecundity.
Fecundity is a component of fitness. The finding that
photoperiod can strongly influence fecundity and hence
fitness has also been shown by Chu and Levin (1989) in
the spionid polychaete Streblospio benedict i. In the case
of N. limnicola. parturition during periods of increasing
daylength appears to select for higher fecundity.
Although both fecundity and life span ofNeanthes lim-
nicola varied with photoperiod, life span varied inversely
with fecundity (Fig. 5). Worms that saw short -*• increas-
ing and long -»• increasing daylength (experiment B) had
a shorter life span in culture but higher fecundity than
the constant short- and constant long-daylength controls
(Fig. 4). Likewise in experiment C, worms in neutral and
neutral -»• decreasing daylengths had the highest fecundity
but the shortest life span. The worm that gave birth to
the most young (232) in all of our experiments lived 224
days, whereas two of the longest-lived worms (564 days
each) produced only 88 and 0 young (Fig. 5). These results
are inconsistent with life history theory that life span,
growth, and fecundity are positively correlated (e.g.. Bell,
1980), but consistent with our previous findings (Fong
and Pearse, 1 992), and indicate that A', limnicola can reach
maximum reproductive potential in half its lifetime.
Consequently, these animals spend the later part of their
lives in maintenance, waiting for a "gate-open period,"
before proceeding with reproduction. This conclusion in-
dicates that the photoperiodic control resulting in season-
ality of parturition in spring-early summer is under strong
selection.
Our experiments reported here demonstrate that in
Neanthes limnicola ( 1 ) parturition occurs with near-max-
imum fecundity in about 8-1 1 months under fixed, neu-
tral photoperiod, (2) parturition is delayed and asynchro-
nous, and fecundity is lowered under short, long, or de-
creasing photoperiods, and (3) parturition is earlier,
synchronous, and only slightly below maximum fecundity
when given seasonally increasing photoperiods after an
initial 3-month exposure to either short, neutral, or long
photoperiod. Moreover, changes in temperature cannot
account for the synchrony displayed by the animals held
296
P. P. FONG AND J. S. PEARSE
under neutral photoperiod (see above). These results from
the neutral phoioperiod treatments suggest that the ani-
mals are programmed to complete their life cycle, from
birth to metamorphosis and parturition, in late winter
and spring. The slight extension (up to 4 months) of the
annual rhythm over the underlying 8-1 1 month endog-
enous rhythm is probably the result of modifying effects
of seasonally changing photoperiod. This situation is sim-
ilar to that described by Olive and Garwood (1983) for
Nereis diversicolor in northern England. Worms main-
tained under constant temperatures and daylengths be-
come sexually mature at the same time as worms in the
field. Oocyte growth in N. diversicolor proceeds at the
same rate at 5°, 10°, 15°, or 20°C, thus the duration of
oogenesis is fixed, and the timing of its completion de-
pends on the time of its initiation. At 5° and 10°C, two
cycles of reproductive activity occur at intervals somewhat
less than 1 year apart. At 1 5°C, all worms become sexually
mature within 1 year of collection. This pattern of repro-
ductive activity suggests an endogenous, gated reproduc-
tive rhythm of circannual periodicity, initiated at birth,
which free-runs for 1-3 years (Olive and Garwood, 1983;
Olive, 1984). No evidence for an exogenous, entraining
zeitgeber has yet been found, however.
Carpet beetles (Anthrenus verbasci) appear to have a
similar endogenous, circannual rhythm of pupation. Bee-
tle larvae maintained in constant darkness at either 22.5°
or 25°C, show one pulse of pupation following their first
winter diapause, then emerge the following spring. But,
larvae held at either 17.5° or 20°C have two peaks of
pupation separated by about 41 weeks (Blake, 1959).
The main component of photoperiod that seems im-
portant for maintaining reproductive synchrony in pop-
ulations ofNeanthes limnicola is increasing daylength, as
is normally experienced in the winter and spring. Not
only did increasing daylength synchronize reproduction
in our experiments reported here, but earlier experiments
showed that the life cycle of A1, limnicola could be shifted
out of phase when the animals were held in seasonally
changing photoperiods out of phase with ambient (Fong
and Pearse, 1992). Similar experiments have shown that
both a fall reproductive diapause in the shrimp Hepla-
carpus pictus (Custer, 1986) and gametogenesis in the sea
star Pisaster ochraceus (Pearse and Eernisse, 1 982; Pearse
et ai, 1986) can be shifted by shifting the phase of the
seasonally changing photoperiods, but these reproductive
cycles remain unaffected when the animals are maintained
under fixed long, neutral, or short daylengths. From the
experiments with P. ochraceus. Pearse et al. (1986) sug-
gested that an underlying endogenous rhythm was syn-
chronized by changing photoperiod. However, unlike N.
limnicola and H. pictus, which live only 1-2 years, indi-
viduals of P. ochraceus live decades or more, spawning
year after year; photoperiodism maintains synchrony
among individuals of P. ochraceus over many years, while
it maintains synchrony within successive generations of
N. limnicola and H. pictus.
In most examples of photoperiodism, including those
of seasonal reproductive activity, one or more "critical
daylengths" appear to trigger events leading to synchro-
nization (Saunders, 1982). The sea urchin Strongylocen-
trotus purpuratus, a marine example, is gametogenic and
full of gametes when under photoperiods less than 12 h,
but gametogenesis is repressed under longer daylengths
(Bay-Schmith and Pearse, 1987). Thus, gametogenesis is
initiated in the fall when daylength drops below 12 h and
slows down in the spring when daylength exceeds 12 h.
Such examples are the basis for "hour-glass," "circadian,"
or similar models explaining photoperiodism (Saunders,
1982; Gwinner, 1986); critical processes require a mini-
mum amount of light each day (hour-glass), or a particular
length of time between two lighted periods (circadian), to
activate a photoperiodic response. However, these models
are inadequate for explaining how changing daylengths,
but not fixed daylengths of any length [(or expected com-
binations such as short -*• long (experiment A)], might
synchronize activities such as parturition in Neanthes
limnicola, diapause in Heptacarpus pictus, or gametoge-
nesis in Pisaster ochraceus (or antler shedding in reindeer,
Goss. 1982, 1984). Rather, the organisms need to be able
to measure daylength and compare it with earlier day-
lengths before initiating a photoperiodic response. As
pointed out by Pearse el al. (1986), new models and in-
sights are needed to explain how changing daylengths can
act to synchronize seasonal activities with an underlying
circannual endogenous rhythm.
Acknowledgments
We thank V.B. Pearse. A.T. Newberry, R. I. Smith, D.
McHugh, and two anonymous reviewers for critiquing
the manuscript. M.E. Steele coordinated the daily lab
checking schedule, and J. Blaney, G. Allison, M. Paddack,
S. Davis, E. Sanford, and D. Ghiglione helped maintain
cultures. Facilities at Long Marine Laboratory were made
available through the Institute of Marine Sciences, Uni-
versity of California, Santa Cruz, and its director Dr. W.
Doyle. This work was supported by graduate student re-
search funds from the Biology Board of Studies, and seed
funds from the Graduate Division, University of Califor-
nia, Santa Cruz; the Society for Sigma Xi; the Dr. Earl
H. and Ethel M. Myers Oceanographic and Marine Bi-
ology Trust; and the Friends of Long Marine Laboratory.
The research was done by the senior author in partial
fulfillment of the requirements for the Ph.D. degree at
the University of California. Santa Cruz.
A PROGRAMMED LIFE CYCLE IN NEANTHES
297
Literature Cited
Bay-Schmith, E., and J. S. Pearse. 1987. Effect of fixed daylength on
the photopenodic regulation of gametogenesis in the sea urchin
Strongylocentrolus purpuratus. Int J Invert Rep. Dev 11: 287-294.
Bell, G. 1980. The costs of reproduction and their consequences. Am.
Sal. 116:45-76.
Blake, G. M. 1959. Control of diapause by an 'internal clock' in An-
threnus verbasci (L) (Col., Dermestidae). Nature 183: 126-127.
Chu, J.-\\ ., and L. Levin. 1989. Photopenod and temperature regu-
lation of growth and reproduction in Streb/ospio henedicli (Poly-
chaeta: Spionidae). Int. J. Invert. Rep Dev. 15: 131-142.
Clark, S. 1988. A two phase photopenodic response controlling the
annual gametogenic cycle in Harmothoe imbricaia (L) (Polychaeta:
Polynoidae). Int. J Invert. Rep. Dev 14: 245-266.
Custer, D. M. 1986. The tidepool shrimp Hepiacarpus pictus: popu-
lation dynamics at Pigeon Point. California and the effects of pho-
toperiod on growth and reproduction. M.Sc. Thesis, University of
California. Santa Cruz. 72 pp.
Kong, P. P. 1991. Environmental control of reproduction in California
Nereidae (Annelida: Polychaeta). Doctoral Dissertation. University
of California, Santa Cruz. 181 pp.
Fong. P. P., and J. S. Pearse. 1992. Photopenodic regulation of par-
turition in the self-fertilizing viviparous polychaete Neanlhes limnicola
from central California. Mar. Biol. 112: 81-89.
Garwood, P. R., and P. J. \V. Olive. 1982. The influence of photoperiod
on ooeyte growth and its role in the control of the reproductive cycle
of the polychaete Harmothoe imbricaia L. Int. J Invert. Rep Dev
5: 161-166.
Goss, R. J. 1982. Control of deer antler cycles by photopenod. Pp. 1-
1 3 in Antler Development in Cervidae, R. D. Brown, ed. Caesar KJe-
berg Wildlife Research Institute. Kingsville, Texas.
Goss, R. J. 1984. Photopenodic control of antler cycles in deer VI.
Circannual rhythms on altered day lengths. / Exp. Zool. 230: 265-
271.
Gvtinner, E. 1986. Circannual Rhythms: Endogenous Annual Clocks
in the Organization oj Seasonal Processes. Springer-Verlag, Berlin.
154pp.
Olive, P. J. \V. 1984. Environmental control of reproduction in Poly-
chaeta. Forlschr. Zool. 29: 17-38.
Olive, P. J. W., and P. R. Garwood. 1983. The importance of long
term endogenous rhythms in the maintenance of reproductive cycles
of manne invertebrates: a reappraisal. Int. J Invert. Reprod. Dev. 6:
339-347.
Olive, P. J. W., and G. Pillai. 1983. Reproductive biology of the poly-
chaete Kefersteinia cirrala Keferstein (Hesionidae). II. The game-
togenic cycle and evidence for photopenodic control of oogenesis.
Int. J Invert. Reprod. Dev. 6: 307-315.
Pearse, J. S., and D. J. Eernisse. 1982. Photopenodic regulation of
gametogenesis and gonadal growth in the sea star Pisasler ochraceus.
Mar. Bio/. 67: 121-125.
Pearse, J. S., D. J. Eernisse, V. B. Pearse, and K. A. Beauchamp.
1986. Photopenodic regulation of gametogenesis in sea stars, with
evidence for an annual calendar independent of fixed daylength. Am.
Zool. 26:417-431.
Saunders, D. S. 1982. Photoperiodism in animals and plants. Pp. 65-
82 in Biological Timekeeping, J. Brady, ed. Cambridge University
Press. Cambridge.
Schierwater, B., and C. Hauenschild. 1990. A photoperiod determined
life-cycle in an oligochaete worm. Biol. Bull. 178: 1 1 1-1 17.
Smith, R. I. 1950. Embryonic development in the viviparous nereid
polychaete, Neanlhes ligliti Hartman. J. Morphol. 87: 414-466.
Reference: Bi«\ Bull 182: 298-304. (June, 1992)
Ultrastructural Study of an Endogenous Energy
Substrate in Spermatozoa of the Sea Urchin
Hemicentrotus pulcherrimus
MASATOSHI MITA'-* AND MASARU NAKAMURA2
^Department of Biochemistry, Teikyo University School of Medicine, Itabashi-kii, Tokyo 173, and
2 Department of Zoology, Faculty of Medicine, Teikyo University, Hachioji, Tokyo 192-03, Japan
Abstract. The morphology of the midpiece in sperma-
tozoa of the sea urchin Hemicentrotus pulcherrimus was
investigated ultrastructurally with particular emphasis on
an endogenous substrate providing energy for motility.
The midpiece was composed of a single toroidal mito-
chondrion surrounding the flagellum. Several lipid bodies
(0.1-0.2 ^m in diameter) were contained in the space be-
tween the mitochondrial outer and inner membranes.
Following incubation with seawater, spermatozoa began
to swim and the lipid bodies became small and finally
disappeared, coincident with a decrease in the level of
phosphatidylcholine (PC), an endogenous substrate for
energy metabolism. In contrast, during incubation in 100
mA/ K+-seawater, in which spermatozoa are immotile.
there was no decrease in the level of PC and the lipid
bodies remained intact. These results strongly suggest that
the PC available for use in energy metabolism is located
in the lipid bodies within mitochondria in the midpieces
of H. pulcherrimus spermatozoa.
Introduction
Spermatozoa are stored for months as immotile cells
in male sea urchins (Gray. 1928; Rothschild, 1959). Upon
spawning in seawater, flagellar movement begins and res-
piration is activated, in close association with Na+-de-
pendent acid extraction (Nishioka and Cross, 1978;
Christen et al ., 1982; Lee et a/., 1983; Bibring el ai, 1984).
Internal alkalization leads to activation of dynein ATPase,
resulting in the inilia:ion of motility (Christen et al., 1983).
The energy for flagellar motility of spermatozoa of the
sea urchin Hemicentmtus pulcherrimus is produced by
Received 28 August 1991: accepted 18 February 1992.
* To whom reprint requests should be addressed.
the oxidation of endogenous phospholipids (Mohri, 1957;
Mita and Yasumasu. 1983a). Similar findings have been
obtained in many other sea urchins, such as Echinus es-
culenlus (Rothschild and Cleland, 1952), Arbacia lixula
(Mohri, 1964), and Strongylocentrotus intermedius (Ko-
zhina et al.. 1978). The spermatozoa of H. pulcherrimus
are generally composed of various phospholipids and
cholesterol (Mita and Ueta, 1988, 1989). Triacylglycerol
(TG) and glycogen are present in trace amounts (Mita
and Yasumasu, 1983a; Mita and Ueta, 1988). The phos-
pholipids include phosphatidylcholine (PC), phosphati-
dylserine, phosphatidylethanolamine, and cardiolipin.
Following incubation with seawater, the level of PC de-
creases, with no change in the levels of other phospholipids
(Mita and Ueta, 1988, 1990; Mita et al.. 1990), indicating
that PC may be a substrate for energy metabolism in sea
urchin spermatozoa. This preferential hydrolysis of PC is
related to the properties of phospholipase A:. The phos-
pholipase A: in //. pulcherrimus spermatozoa has high
substrate specificity for PC (Mita and Ueta, 1990), which
may therefore be used specifically for energy metabolism.
Recently. PC has been shown to be abundant in H.
pulcherrimus sperm midpieces (Mita et ai. 1991). Fol-
lowing the initiation of motility, the PC content of sperm
midpieces decreases significantly, while that in sperm
heads and tails does not change (Mita et al., 1991). Phos-
pholipase A2 activity is also distributed in the midpieces
(Mita et al.. 1991). Thus, PC available for use in energy
metabolism is located in the midpieces. It has also been
reported that the midpieces of Brissopsis lyrifera (Afzelius
and Mohri. 1966) and Echinarachinus parma (Summers
and Hv lander, 1974) contain a single mitochondrion and
lipid globules. The lipid globules are spherical and located
in the posterior region between the base of the mitochon-
298
MIDP1ECES OF SEA URCHIN SPERM
299
drion and the plasma membrane (Afzelius and Mohri.
1966). Although similar lipid globules have not been ob-
served in spermatozoa of other sea urchin species, it has
been reported that lipid bodies are present in A. punctulata
(Longo and Anderson, 1969) and A. lixitla (Cosson and
Gulik, 1982) spermatozoa. The lipid body differs from
the lipid globules, because the former is located inside the
mitochondrion and it is relatively smaller than lipid glob-
ules (Longo and Anderson, 1969; Cosson and Gulik,
1982). In the present study, the midpieces of H. pulcher-
rimus spermatozoa were examined ultrastructurally to
clarify further the energy metabolism of sea urchin sper-
matozoa.
prefixed in 2.5% glutaraldehyde ASW solution for 40-60
min at 4°C; a volume of sperm suspension was mixed
with the same volume of cold 5% glutaraldehyde in 80%
ASW. The prefixed spermatozoa were rinsed with cold
ASW and post-fixed with 1% OsO4 in ASW for 2 h at
4°C. Samples were washed in distilled water, and then
immersed in saturated aqueous uranyl acetate for 1 h for
block staining. After dehydration in a graded series of
ethanol solutions, the specimens were embedded in epoxy
resin and ultrathin sections were cut on a Reichert Ultra-
cut ultramicrotome. After staining the specimens with lead
citrate, we used a Hitachi 7000 or JEM 100 CX electron
microscope to observe them.
Materials and Methods
Materials
Spawning of stored spermatozoa of the sea urchin H.
pii/c/ierrimus was induced by injecting 0.5 M KC1 into
the coelomic cavity. Semen was always collected freshly
as "dry sperm" and kept undiluted on ice.
Incubation of spermatozoa
Dry sperm were diluted 100-fold in artificial seawater
(ASW) consisting of 458 mA/NaCl, 9.6 mM KC1, 10 mM
CaCl:, 49 mM MgSO4, and 10 mM Tris-HCl, pH 8.2.
After dilution and incubation at 20°C, the sperm suspen-
sion was centrifuged at 3000 X g for 5 min at 0°C. In 100
mM K+-seawater, Na+ was substituted for K+.
Determination of PC concentration
Total lipids were extracted from spermatozoa using the
method of Bligh and Dyer (1959). PC levels were deter-
mined by high-performance thin-layer chromatography.
as described previously (Macala et ai, 1983; Mita and
Ueta, 1988). PC content consumed during incubation for
1 h was calculated from the absolute value of PC before
and after incubation.
Oxygen consumption
Oxygen consumption in a sperm suspension was mea-
sured polarographically with an oxygen consumption re-
corder (MD-1000. lijima Electronics MFG Co., Japan).
Twenty-five n\ of dry sperm were incubated in 2.5 ml of
ASW in the closed vessel of the oximeter at 20°C.
Preparation for electron microscopy
Dry sperm were diluted 100-fold in ASW and incubated
at 20°C. At appropriate intervals, the spermatozoa were
Results
In longitudinal sections through spermatozoa of H.
pulcherrimus. the midpiece was observed to consist of a
single toroidal mitochondrion (Fig. 1). The midpiece did
not contain the lipid globules observed in the spermatozoa
of B. lyrifera (Afzelius and Mohri, 1966) and E. parma
(Summers and Hylander, 1974). A region between the
mitochondrial outer and inner membranes — intramem-
brane space — was dilated in a band nearest the flagellum
Figure 1. Longitudinal section (a) and schematic representation (h)
of a spermatozoon of llemicentrotus pulcherrimus Arrow heads show
lipid bodies (LB). C: proximal centriole, F: flagellum, G: acrosomal gran-
ule, M: mitochondrion. N: nucleus, o.m.: mitochondnal outer membrane,
p.m.: plasma membrane. SF: subacrosomal fossa, i.s.: intramembrane
space. XI 9,700.
300
M. MITA AND M. NAKAMURA
2a
Figure 2. Longitudinal (a) and transverse (b) sections through the mitochondrial region of spermatozoa
before incubation in seawater. F: tlagellum, i.m.: inner mitochondnal membrane, i.s.: intramembrane space,
LB: lipid body, M: mitochondrion. N: nucleus, o.m.: outer mitochondrial membrane, p.m.: plasma membrane.
X58.800.
Figure 3. Electron micrograph of spermatozoa before incubation with seawater. Arrow heads show lipid
bodies, xl 1.800.
MIDPIECES OF SEA URCHIN SPERM
301
and contained low-electron-density lipid bodies (Fig. 2).
These lipid bodies were irregular in profile and about 0.1-
0.2 ^m in diameter. All of the spermatozoa in semen con-
tained the lipid bodies within their mitochondria (Fig. 3).
The same lipid bodies were also observed in spermatozoa
present in the testis (data not shown).
When dry sperm were diluted and incubated in ASW,
spermatozoa began to swim and the amount of sperm PC
decreased (Fig. 4). About 6 ^g of PC was consumed in
10* spermatozoa following incubation for 1 h (Table I).
In addition to PC consumption, respiration was activated
and about 0.27 j/mol O:/h/104 spermatozoa was con-
sumed. These findings confirm the previous observations
(Mita and Ueta, 1988, 1990). Longitudinal and transverse
sections of the midpieces of spermatozoa were examined
following incubation in ASW. After 5 min of incubation,
changes were noted in the structure of the inner ring of
the mitochondrion. Although lipid bodies were still pres-
ent, they had shrunk. In addition, a gap was observed to
have opened between the plasma membrane and the mi-
tochondrial outer membrane (Fig. 5b, e). After 30 min of
incubation, the inclusion bodies and the inner ring of the
mitochondrion had disappeared (Fig. 5c, f). Various
structural features of the mitochondrion, such as the
number of cristae and the thickness of the membranes,
did not change during incubation in ASW.
Because sea urchin spermatozoa incubated in high K+-
seawater are immotile and their respiration extremely low
(Schackmann ct a/., 1981; Mita and Yasumasu, 1983b,
1984), the effect of a high-K+ environment on the lipid
bodies of the midpiece was examined. After incubation
in 100 mM K+-seawater for 1 h at 20°C. neither oxygen
nor PC was consumed by the spermatozoa (Table I), and
the lipid bodies of the midpiece remained intact (Fig. 6a,
b). Thus, the disappearance of the lipid bodies was cor-
related with the decrease in the level of PC.
Table I
Phosphatidylcholine and oxygen
in \cti urchin srtcrmatiwa
40
a
S35
o
— 30
'c
0)
§25
a
20
0 15 30 45 60
Incubation time (min)
Figure 4. The change in level of phosphatidylcholine (PC) in sea
urchin spermatozoa following incubation in seawater. Each value is the
mean of four separate experiments. Vertical bars show S.E.M.
Conditions
PC consumption
fcg/h/109
sperm)
O2 consumption
(fimol O2/h/10g sperm)
Seawater
100 mM KH -seawater
6 ± 1
N.D.
0.27 ± 0.02
<0.01
Dry sperm were diluted 100-fold in either seawater or 100 mA/ K+-
seawater and incubated for 1 h at 20°C. Values are means ± S.E.M.
obtained from four separate experiments. N.D., not detectable.
Discussion
The present study demonstrated lipid bodies in the in-
tramembrane space of the mitochondrion in the sperm
midpiece of// pnlc/icmmus (Fig. 1). Following incuba-
tion of spermatozoa in seawater, these lipid bodies dis-
appeared gradually (Fig. 5). although they still remained
after incubation in 100 mM K+-seawater (Fig. 6). These
observations were correlated with changes in the level of
intracellular PC (Fig. 4). suggesting that PC available for
use in energy metabolism is related to the lipid bodies
within the mitochondria of the midpiece. Similar lipid
bodies have been observed in the spermatozoa of A. punc-
tulata (Longo and Anderson, 1969) and A. lixula (Cossin
and Glik, 1982). It has also been reported that .-1. lixula
spermatozoa obtain energy for movement from the oxi-
dation of endogenous phospholipid (Mohri. 1964). These
findings also support the hypothesis that the lipid bodies
within mitochondria are reservoirs of endogenous PC
substrate in sea urchin spermatozoa.
We also showed that about 6 ^g of PC was consumed
in 109 spermatozoa following incubation for 1 h (Fig. 4,
Table I). Because this amount was only '/5 of the total PC,
the remaining % of cellular PC may be membrane-bound
and therefore inaccessible as an energy substrate for mo-
tility. About 0.27 jumol O2/h/109 spermatozoa were con-
sumed (Table I). This degree of oxygen consumption is
enough to account for the consumed PC, as mentioned
previously (Mita and Yasumasu, 1983; Mita et ai. 1990).
Presumably, the fatty acid liberated from PC in the lipid
bodies is metabolized through /i-oxidation to produce ATP.
Unfortunately, there is little direct evidence to indicate
whether the content of the lipid bodies is, in fact, PC. A
cytochemical study would be useful to identify PC in the
lipid bodies, although an antibody against PC would be
difficult to prepare because PC is a common membrane
component. We are now investigating the role and char-
acteristics of the lipid bodies to provide useful insights
302
M. MITA AND M. NAKAMURA
Figure 5. Longitudinal (a-c) and transverse (d-f) sections through the mitochondria] region of spermatozoa
before (a. d) and after incubation in seawater for 5 min (b, e) and 30 min (c. f). Arrow heads show lipid
bodies. F: flagellum. M: mitochondrion. N: nucleus. X42.500.
into the direct mechanism of energy metabolism in sea
urchin spermatozoa.
In contrast to the PC used in H. pulcherrimus, Glyp-
tocidaris cremtlaris spermatozoa use TG as a substrate
for energy metabolism (Mita. 1991). There are several
lipid globules at the bottom of the midpiece in G, cren-
itlaris spermatozoa (Mita and Nakamura, 1992), similar
to those in the spermatozoa of B. lyrifera (Afzelius and
Mohri, 1966) and E. parnia (Summers and Hylander,
1974). After incubating G. crenularis spermatozoa with
seawater, both the number and the size of the lipid globules
decreased, coincident with a decrease in the TG level.
M1DPIECES OF SEA URCHIN SPERM
303
Figure 6. Longitudinal (a) and transverse (b) sections through the mitochondrial region of spermatozoa
after incubation in 100 m.U K+-seawater for 1 h. Arrow heads show lipid bodies. F: flagellum. M: mito-
chondrion, N: nucleus. X42.500.
However, neither TG (Mita and Ueta, 1988) nor lipid
globules (Fig. 1) are present in H. pulcherrimus sperma-
tozoa. Thus it appears that TG is related to the lipid glob-
ules.
Acknowledgments
The authors are grateful to Dr. N. Ueta, Teikyo Uni-
versity School of Medicine, for his encouragement and
valuable advice, and to Dr. N. Usui. Teikyo University
School of Medicine, and Dr. V. Nagahama. National In-
stitute for Basic Biology, for their valuable comments.
Thanks are also extended to Dr. S. Nemoto and the staff
of the Tateyama Marine Laboratory, Ochanomizu Uni-
versity, for their assistance in collecting the sea urchins.
This work was supported in part by a Grant-in-Aid
(03740396) from the Ministry of Education, Science and
Culture of Japan.
Literature Cited
Afzelius, B. A., and II. Mohri. 1966. Mitochondria respiring without
exogenous substrate: a study of aged sea urchin spermatozoa. Exp.
Cell Res. 42: 10-17.
Bibring, I. J., J. Baxandall, and C. C. Harter. 1984. Sodium-dependent
pH regulation in active sea urchin sperm. Dcv. Biol. 101: 425-435.
Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid
extraction and purification. Can. J. Biochem. Phrsiol 37: 91 1-917.
Christen, R., R. \V. Schackmann, and B. M. Shapiro. 1982. Elevation
of intracellular pH activates sperm respiration and motility of sperm
of the sea urchin Strongylocentrotus purpuratus. J Biol. Clicm 257:
14.881-14,890.
Christen, R.. R. \V. Schackmann, and B. M. Shapiro. 1983. Metabolism
of sea urchin sperm. Interrelationships between intracellular pH.
ATPase activity, and mitochondnal repiration. J Biol. Chem. 258:
5392-5399.
Cosson, M. P., and A. Gulik. 1982. Description of the mitochondria-
axoneme junction in sea urchin spermatozoa: presence ol a flagellar
necklace. J Vltrastntcl. Res. 79: 47-57.
Gray, J. 1928. The effect of dilution on the activity of spermatozoa.
J Exp. Biol. 5: 337-344.
Kozhina, V. P., T. A. Terekhova, and V. I. Stretashev. 1978. Lipid
composition of gametes and embryos of the sea urchin Strongvh-
cenlrolux inlermedius at early stages of development. Dev. Biol 62:
512-517.
Lee, H. C., C. Johnson, and D. Epel. 1983. Changes in internal pH
associated with initiation of motility and acrosome reaction of sea
urchin sperm. Dcv. Biol. 95: 31-45.
Longo, F. J., and E. Anderson. 1969. Sperm differentiation in the sea
urchins Arbacia /ndiclulcila and Strogylocentrotus purpuratus. J.
Ullruslrucl. Ki-.v. 27: 486-509.
Macala, L. J., R. K. Yu, and S. Ando. 1983. Analysis of brain lipids
by high performance thin-layer chromatography and densitometry.
J. Lipid Rex 24: 1243-1250.
Mita, M., and I. Yasumasu. 1983a. Metabolism of lipid and carbo-
hydrate in sea urchin spermatozoa. Gamete Res. 7: 133-144.
Mita, M., and I. Yasumasu. 1983b. Effect of Na*-free seawater on
energy metabolism in sea urchin spermatozoa with special reference
to coenzyme A and carmtine derivatives. Gamete Res 7: 259-267.
Mita, M., and 1. Yasumasu. 1984. The role of external potassium ion
in activation of sea urchin spermatozoa. Dev. Growth Differ. 26: 489-
495.
Mita, M., and N. Ueta. 1988. Energy metabolism of sea urchin sper-
matozoa, with phosphatidylcholine as the preferred substrate.
Biochiin. Biophy\. Ada 959: 361-369.
Mita, M., and N. Ueta. 1989. Fatty chain composition of phospholipids
in sea urchin spermatozoa. Comp. Biochem. Physiol. 92B: 319-322.
Mita, M., N. Ueta, I . Harumi, and N. Suzuki. 1990. The influence of
egg-associated peptide on energy metabolism in sea-urchin sperma-
tozoa: the peptide stimulates preferred hydrolysis of phosphatidyl-
choline and oxidation of fatty acid. Bwchim. Biophys. Actu 1035:
175-181.
Mita, M., and N. Ueta. 1990. Phosphatidylcholine metabolism for en-
ergy production in sea urchin spermatozoa. Biochim. Biop/iys. Ada
1047: 175-174.
Mita, M., T. Harumi, N. Suzuki, and N. Ueta. 1991. Localization and
characterization of phosphatidylcholine in sea urchin spermatozoa.
/ Biochem. 109: 238-242.
304 M. MITA AND M. NAKAMURA
Mita, M. 1991. Energy metabolism of spermatozoa of the sea urchin Rothschild, Lord, and K. \V. Cleland. 1952. The physiology of sea-
Glyptocidans ci'cii\il,iri\. Mol. Reprod. Dev. 28: 280-285. urchin spermatozoa. The nature and location of the endogenous sub-
Mila, M., and M. Nakamura. 1992. Lipid globules at the midpieces strate. J. Exp. Biol. 41: 66-71.
of 6/r • ii'iiiilaris spermatozoa and its relation to energy Rothschild, Lord. 1956. The physiology of sea-urchin spermatozoa.
metabol" m. Mol. Reprod. Dev. (in press). Action of pH, dinitrophenol, dinitrophenol + versene, and usnic
Mohri, H. 1957. Endogenous substrates of respiration in sea-urchin acid or O2 uptake. / Exp. Biol. 33: 155-173.
spermatozoa. J. Fac. Set. Tokyo Univ. IV. 8: 51-63. Schackmann, R. \V., R. Christen, and B. M. Shapiro. 1981. Membrane
Mohri, H. 1964. Phospholipid utilization in sea-urchin spermatozoa. potential depolarization and increase intracellular pH accompanying
Puhh. Sla;. Zoo/. Napoli 34: 53-58. the acrosome reaction of sea urchin sperm. Proc. Nail. Acad. Sci.
Nishioka, D., and N. Cross. 1978. The role of external sodium in sea L'SA 78: 6066-6070.
urchin fertilization. Pp. 403-4 1 3 in Cell Reproduction: In Honor of Summers, D. G., and B. L. Hylander. 1974. An ultrastructural analysis
D. Mazia. E. R. Dirksen, D. M. Prescott, and S. F. Fox, eds. Academic of early fertilization in the sand dollar, Echinarachnius parma. Cell
Press. New York. Tissue Res 150: 343-368.
Reference: «/«>/. Bull. 182: 305-323. (June. 1992)
Studies on the Structure and Function of the Larval
Kidney Complex of Prosobranch Gastropods
BRIAN R. RIVEST
Department of Biological Sciences, Stale University of New Yi>rl< at Cortland,
Cort land. New York 13045
Abstract. The larval kidneys of prosobranch gastropods
have long been assumed to be involved in handling wastes,
but with little supporting evidence. In this study, the larval
kidneys of Searlesia dira and Niicella canalicu/ata were
studied with light, electron, and fluorescence microscopy.
They consist of three cell types: ( 1 ) a large external ab-
sorptive cell swollen with heterophagosomes and pos-
sessing an endocytotically active external surface; (2) an
internal crystal cell with numerous vacuole-bound crystals
of a calcium salt and with morphologically complex can-
aliculi: (3) an internal pore cell characterized by slit-pores
that lead to subsurface cisternae. a tubular network, and
one or two ciliated ducts that open into the hemocoel.
Empirical evidence indicates that the absorptive cell rap-
idly takes up and stores albumen proteins from the cap-
sular fluid. Absorptive cells were found in 1 7 of 1 9 species
tested, representing three prosobranch orders, but were
not found in 2 opisthobranch or 1 pulmonate species. We
hypothesize that the absorptive cells have become spe-
cialized for the uptake of capsular albumen prior to the
functional differentiation of the gut. However, the nutri-
tional importance of the absorbed albumen proteins and
the functions of the crystal and pore cells are presently
unknown. No evidence for an excretory function was
found for the larval kidney complex: it may be a vestigial
protonephridium. the components of which have become
disorganized and functionally altered.
Introduction
Larval kidneys are prominent structures found in the
neck region in many prosobranch embryos. They have
long been considered to excrete actively or to accumulate
waste products (Bobretzky, 1877; Heymons, 1893;Glaser.
Received 30 December 1991: accepted 17 March 1992.
1904: Pelseneer, 1911: Franc. 1940). The hypothesis that
these structures are organs of accumulation is attractive
because it explains the large cells as an adaptation for
encapsulated development (Pelseneer, 1911; Eisawy and
Sorial. 1974). Embryos might better sequester their wastes
than pollute the surrounding capsular fluid.
Prosobranch larval kidneys have been described as
paired, laterally located uni- or multicellular structures
that protrude from the region behind the velum, and are
usually considered to be of ectodermal origin (Casteel,
1904; Portmann. 1930: Franc, 1941). Bobretzky (1877),
Conklin (1897). Glaser (1904), and D'Asaro (1966) de-
scribed them as enlarged ectodermal cells that either re-
leased excretory granules or were completely cast offbefore
hatching. In Ocenehru ueicitlala. Franc ( 1940) found the
larval kidneys to consist of a larger outer vacuolated cell
overlying a smaller internal cell containing green vacuoles.
He tried unsuccessfully to demonstrate experimentally an
excretory function of the larval kidneys. However, he did
find what he described as yolk platelets in the larger ex-
ternal cell of O. ac/culata and Thais haemastoma (Franc,
1940. 1941). Although Franc realized that the presence
of yolk would suggest that larval kidneys were involved
in something other than excretion, he retained the classical
assumption that they were organs of waste accumulation
and hypothesized that the yolk supplied the energy used
for waste acquisition (Franc, 1941 ).
Although the presence of larval kidneys in prosobranch
embryos has been reported by numerous authors, they
are poorly understood. The material accumulated in the
enlarged ectodermal cells of the larval kidneys has not
been identified, although Glaser (1904) claimed that an
aqueous extraction of larval kidney cells contained dilute
urea. Furthermore, little is known of the ultrastructure of
prosobranch larval kidneys. In this study I describe the
ultrastructure of the three cells that make up the larval
305
306
B R RIVEST
kidney in the marine snail Searlesia dim and the ontogeny
of these cells, from the early trochophore to the mid-veliger
stage. Similar observations of the larval kidneys of Nucella
cana/icii/a/ti are included. Morphological and experi-
mental evidence shows that the large size of the larval
kidneys is due to endocytotically absorbed albumen pro-
teins, not stored waste products. Twenty other species of
gastropods were tested for larval kidneys that absorb dis-
solved proteins from the surrounding fluid.
Materials and Methods
Collection of egg masses and egg capsules
Most egg masses or capsules used in this study were
collected in intertidal or subtidal areas around San Juan
Island. Washington, or were obtained from animals
maintained in aquaria at the Friday Harbor Laboratories.
Egg capsules ofOcenebrajaponica were collected on oyster
flats in southern Puget Sound. Washington, and those of
Nucellu lapil/us at York Beach, Maine.
Specimen preparation for light and electron microscopy
Searlesia dira and Nucella canaliculala embryos to be
examined with scanning electron microscopy were re-
moved from their capsules, rinsed in filtered seawater,
relaxed with 7.5% magnesium chloride and fixed for one
to two hours at room temperature in 2% osmium tetroxide
in 1 .25% sodium bicarbonate at pH 7.2 (Wood and Luft,
1965). The embryos were then rinsed in distilled water,
dehydrated in ethanol, critical point dried, mounted, and
coated with gold. The embryos were examined with an
ETEC Autoscan or a JEOL JSM-35 scanning electron
microscope.
Satisfactory fixation of larval kidney cells for trans-
mission electron microscopy was difficult. Fixatives that
appeared suitable for adjacent ectoderm and subjacent
endoderm often produced fixation artifacts in the larval
kidney, particularly in the outer absorptive cell. The het-
crophagosomes of this voluminous cell appeared partic-
ularly sensitive to the osmolarity of the fixative. The larval
kidneys of Searlesia dira were most satisfactorily fixed in
a cacodylate buffered glutaraldehyde solution containing
ruthenium red. a method modified from Cavey and Clo-
ney (1972). Embryos removed from their capsules and
rinsed with seawater were placed in 2% glutaraldehyde in
a solution buffered by 0.2 M sodium cacodylate adjusted
to 1000-1 100 mOsM with sucrose and containing 0.05%
ruthenium red and 0.002 M calcium chloride. Final pH
was adjusted to 7.4. This fixative did not work well for
the larval kidneys of Nucella canalicu/ata. which showed
better preservation using a 3% glutaraldehyde solution
containing 0.1 M sodium cacodylate and 0.001 /I/calcium
chloride and adjusted to 1000-1 100 mOsM with sodium
chloride. This worked satisfactorily for light microscopy
but produced severe artifacts at the ultrastructural level.
The embryos were postfixed in 2% osmium tetroxide
in freshly mixed 1.25% sodium bicarbonate (pH 7.2) for
one hour at room temperature. The embryos were then
dehydrated and embedded in EPON (Luft, 1961).
One micrometer thick sections for light microscopy
were stained with a mixture of Azure II and methylene
blue in 0.5% sodium borate (Richardson et ai, 1960).
Serial thin sections for transmission electron microscopy
were stained with uranyl acetate and lead citrate (Reyn-
olds, 1963) and examined with a Philips EM-300 electron
microscope.
The Searlesia dira developmental stages that were sec-
tioned included early trochophores prior to or just after
the initiation of nurse egg feeding, late trochophores near
or just after the end of nurse egg feeding, early veligers
with the cephalopedal elements elevated from the main
embryonic body and the mantle fold near the anterior
end, and mid-veligers with greater differentiation of the
cephalopedal elements and a small mantle cavity. The
Nucella canaliculata developmental states that were sec-
tioned included late trochophores, early veligers, and mid-
veligers.
Exposure of embryos to test solutions
To study the uptake of material by the larval kidneys,
embryos were initially exposed to isosmotic solutions of
ferritin. Later experiments included placing embryos in
isosmotic solutions containing fluorescein, bovine serum
albumin (BSA). capsular albumen, fluoresceinisothio-
eyanate labelled BSA (FITC-BSA, Sigma Chemical Com-
pany), and/or FITC-capsular albumen.
The capsular albumen proteins of Searlesia dira were
labelled with FITC as follows: ( 1 ) seventy capsules less
than three weeks old (before the embryos had begun to
feed on the nurse eggs) were opened in 4°C filtered sea-
water containing 10 Mg/ml of the antibiotic rifampicin.
The albumen was separated from the embryos and nurse
eggs by using a 152 ^m Nitex screen. (2) The seawater-
diluted albumen was centrifuged to remove any debris.
(3) The supernatant was placed in dialysis tubing on a
bed of Aquacide (CalBiochem) to extract water and con-
centrate the proteins. (4) The protein content of the so-
lution was estimated by spectrophotometrically measuring
absorbance at 280 m// and assuming the extinction coef-
ficient of capsular albumen is 6.2 (intermediate between
the extinction coefficients of BSA and human serum al-
bumin). (5) Sufficient NaHCO3 was added to make a 0.1
M solution with a pH of about 9.2. (6) FITC (approxi-
mately one-tenth the weight of the estimated protein in
the solution) was dissolved in one drop of 0.5 A' NaOH.
then the buffered albumen solution was added. (7) The
PROSOBRANCH LARVAL K1DMYS
307
Figure 1. SEM of a Scarlcxia dira embryo shortly after the onset of feeding on nurse eggs showing the
protruding absorptive cells of the larval kidneys.
Figure 2. SEM of an older feeding stage o( S. dira fixed while swallowing a nurse egg. The absorptive
cells are larger and protrude further from the embryo's surface.
Figure 3. SEM of an 5. dira intracapsular veliger showing the developing cephalopedal structures and
the bulbous nature of the left absorptive cell.
Scale bars represent 100 ^m; ac. absorptive cell of the larval kidney; cv, cephalic vesicle: f, foot; oc, oral
cilia; m. mantle margin; ne. nurse egg; v. velum.
mixed solutions were incubated at 20°C for 20 h. (8) Un-
reacted FITC was removed by adding 50 mg powdered
charcoal per mg FITC. spinning the charcoal down after
30 min and repeating this step. (9) The pH was adjusted
to that of seawater with 0.1 A' HC1 and the final solution
was dialyzed against seawater for four hours. This solution
contained approximately 0.37 mg/ml protein, as deter-
mined spectrophotometrically.
Embryos were transferred to a test solution at 12°C
after first being removed from their capsules and rinsed
in filtered seawater. After periods of 10 min to 4 h. the
embryos were rinsed again in seawater and examined us-
ing light, fluorescence, or transmission electron micros-
copy.
To examine the nature of the crystals found in one of
the inner cells of Searlesia dim larval kidneys, live and
heat-killed embryos were examined with polarized light
while exposed to one of the following solutions: Millonig's
phosphate buffer at pH 3.5 to 7.5: 0.1 .17 glycine at pH
9.2: buffered (pH 8.5) and unbuffered (pH 3.0) 10% for-
malin in seawater: 0.1 M EGTA (pH 9.0) in seawater,
100% ethanol: 10% Triton-X 100 in distilled water; sat-
urated aqueous calcium chloride. Some live embryos were
placed in a 2cr alizarin red S solution in seawater for 48 h.
Some Searlesia dira embryos at the early veliger stage
of development were micro-injected in the hemocoel be-
hind the velum with an isosmotic ferritin solution. After
10 to 60 min. these embryos were fixed and prepared for
examination with transmission electron microscopy.
Larva/ kidney histochemistry
The periodic acid-Schiff reaction was performed on 10
^m thick serial paraffin sections of Searlesia dim embryos
fixed in Carnoy's fixative (Humason, 1972). As a control
for glycogen. some sections were incubated in saliva at
37 °C for 1 h before applying the PAS technique. The
hexamine silver method (Pearse, 1972) was used to test
for urates in paraffin sections of embryos fixed in 100%
ethanol or Carnov's fixative.
Results
Searlesia dira
External morphology />/ the larval kidneys. The two
larval kidneys develop laterally on the embryo just pos-
terior to the mouth (Fig. 1 ). They are discernible as pro-
trusions on live embryos within four weeks of oviposition
at 12°C. before the embryos have begun to feed on nurse
eggs. At this early trochophore stage, the larval kidneys
are circular in outline and 40 to 50 ^m in diameter. During
the one to two and a half weeks that the embryos are
feeding on nurse eggs and for a week or two afterwards.
the larval kidneys enlarge, bulging outward from the em-
bryo (Fig. 2). As they enlarge they become more hemi-
spherical, but their shape and size varies among embryos.
After nurse egg feeding has ended, the cephalopedal
elements grow out from the previously roundish embryo.
At this point the larval kidneys are on the neck region of
these veligers, posterior to the developing velar lobes but
anterior to the forming mantle cavity. The right larval
kidney has typically become elongate in the dorso-ventral
axis, extending dorsally from the midlateral line. In con-
trast, the left larval kidney has become more bulbous,
protruding at the level of, or slightly ventral to. the left
lateral midline (Fig. 3).
308
B. R. RIVEST
Capsular fluid
mv
dbN!
Haemocoel oS^-tn^
sc sp
.
Figure 4. Schematic drawing of a section through the three cells making up a larval kidney complex in
a Searlfsia (lira veliger. Proportions are not relative. The crystals in the vacuoles of the crystal cell dissolve
during fixation and were not seen in sections.
Figure 5. One micrometer thick section of the larval kidney complex in S (lira The pore cell can only
be positively identified in fixed material with TEM.
Scale bar represents 25 jim; ac. absorptive cell; ca. canaliculus; cc. crystal cell; cd, ciliated duct (one of
two); db, dark body; ec, ectoderm; el, external lamina; en. endoderm; ev, endocytotic invaginations and
vesicles: h. heterophagosomes; ma, macromere of swallowed nurse egg; mv. microvilli; pc, pore cell; tn.
tubular network; sc, subsurface cisternae; sp, split pores; v, vacuoles.
The larval kidneys have achieved their maximum size
by the time the advancing mantle margin begins to form
the mantle cavity. The elongate right larval kidney can
be up to 100 /urn long and 35 /jm wide, while the more
spherical left larval kidney may reach 80 /^m in diameter.
The larval kidneys have no distinctive coloration, but ap-
pear more yellowish-white than adjacent ectodermal cells.
Both larval kidneys maintain their relative shape, position,
and size during the intracapsular veliger phase of devel-
opment. During the intracapsular metamorphosis that
follows, the larval kidneys diminish in size until they are
no longer discernible before the end of velar lobe resorp-
tion. There is no evidence suggesting that the larval kid-
neys fall off or spontaneously release their contents into
the lumen of the capsule. There is also no evidence of a
pore near the larval kidney as reported in Ilyanassa ob-
soleta (Tomlinson. 1987).
Ultrastructure ol ilie larva/ kidney* ofSearlesia dira.
Each larval kidney is a complex of three cells: absorptive,
crystal, and pore cells. An exception was found in one
sectioned embryo that had a small second crystal cell ad-
jacent to the main one. Absorptive cells are named after
their ability to absorb external proteins. Crystal cells have
been reported in prosobranch larval kidneys by Portmann
( 1930) and Fioroni ( 1985). Larval kidney pore cells have
surface specializations similar to pore cells found in mol-
luscan connective tissue (Sminia and Boer, 1973), al-
though they are called rhogocytes by Fioroni el a/. (1984).
Each cell type is identifiable in thin sections of the pre-
feeding trochophores, the earliest stage examined. Figure
4 schematically illustrates the distinctive features of these
three cells in an early veliger.
Absorptive cell. The absorptive cell is the outermost
and largest of the three cells that comprise the larval kidney
complex (Figs. 4. 5). The external surface of the absorptive
cell is exposed to the capsular fluid surrounding the em-
bryo. Morphological and experimental evidence indicate
that it is a surface active in receptor-mediated endoc\ tosis.
It is elaborated into numerous microvilli up to 1.8 /urn
long that have endocytotic vesicles forming around their
bases (Fig. 6). Adjacent endocytotic vesicles apparently
fuse shortly after their formation, and the resulting small
heterophagosomes fuse with larger, deeper heterophago-
somes. The size of the absorptive cell may be due mostly
to a few large heterophagosomes, which can reach 35 ^m
in diameter. In live embryos, large heterophagosomes can
easily be detected in the absorptive cell with light mi-
croscopy, they appear refractile and colorless.
PROSOBRANCH LARVAL KIDNEYS
309
When embryos are soaked in a territin solution for 10
min or more, ferritin is found in endocytotic vesicles and
heterophagosomes (Fig. 6). Thus, the heterophagosomes
contain material brought into the cell by endocytosis. In
untreated embryos, the contents of the heterophagosomes
appear fairly uniform with scattered granules, fibrils, or
membranous structures. The density of the contents varies
among the heterophagosomes within a cell. This appears
to be due to differences in concentration of the contents,
not to a segregation of material brought into the cell. This
is supported by the fact that ferritin is eventually found
in all the heterophagosomes in treated embryos. After an
embryo has soaked 20 min in a reddish ferritin solution
at 15-18°C, the absorptive cell takes on a light reddish
color. This color is restricted to small heterophagosomes
in the external cortex of the cell. After another 30 min in
the solution, the color can be perceived in larger hetero-
phagosomes deeper inside the cell. After an hour or more,
the absorptive cells are deep red. This contrasts with the
rest of the ectoderm which has remained translucent and
colorless. There is also no concentration of ferritin in cells
lining the gut.
The heterophagosomes comprise an estimated 80-90%
of the absorptive cell's volume when the cell has reached
its maximum size. The nucleus is usually centrally located,
surrounded by heterophagosomes that indent the nuclear
membrane. Few organelles occur in the apical cytoplasm;
Golgi cisternae, mitochondria, and dense concentrations
of rough ER are found laterally and basally. The ground
cytoplasm appears more electron-dense than that of ad-
jacent ectodermal cells or the subjacent crystal cell (Fig.
7. upper right).
The heterophagosomes showed a positive periodic-acid
Schiff reaction which did not change after incubation of
sections with saliva. This indicates the presence of a car-
bohydrate or carbohydrate-protein complex, but not of
glycogen (Humason, 1972). Also, glycogen was not seen
in electron micrographs. Tests for urates were negative,
and no crystals or large granules were found in the het-
erophagosomes or cytoplasm. The heterophagosomes
contained a few myelinic bodies, but these were uncom-
mon and probably fixation artifacts. Experimental evi-
dence (described below) indicates that the heterophago-
somes contain albumen proteins taken in from the cap-
sular fluid.
Crystal cell. The crystal cell is subjacent to the absorp-
tive cell (Figs. 4. 5, 7). The plasmalemmae of these two
cells are closely apposed. although occasionally small in-
tercellular spaces are found. Junctional complexes be-
tween them were not found. The crystal cell covers most
of the inner surface of the absorptive cell.
In live embryos, the crystal cell can be discerned under
the absorptive cell because of its crystals. These crystals
can be seen most easily in early veliger stages. They are
irregularly shaped and usually occur singly within vacu-
oles, but may be in clusters of up to five crystals. The
crystals measure up to 5 //m in diameter and are birefrin-
gent under polarized light. There are from 20 to 60 crystals
in each crystal cell in early veligers. The crystals are sol-
ubilized, as determined by the loss of birefringence, within
2-5 min in heat killed embryos in pH 3.5 Millonig's
phosphate buffer or in pH 9.0 0.1 M EGTA in seawater.
They quickly dissolved when live embryos were fixed in
unbuffered pH 3.0 10% formalin, but not in buffered pH
8.5 formalin. The crystals were not solubilized within 10
min in phosphate buffers with a pH from 5.0 to 7.5, in
pH 9.2 0.1 M glycine, in ethanol, or in live embryos in
the EGTA solution. The crystals quickly dissolved in 1
TV NaOH and in a 10% solution of the detergent Triton-
X 100 as soon as the vacuoles containing the crystals broke
open. However, in 10% Triton-X 100 made up in a sat-
urated calcium chloride solution, the crystals dissolved
more slowly after the vacuoles were lysed. In live embryos
that were soaked in a filtered 2% alizarin red S solution
in seawater for 24 h, the normally lightly greenish-yellow
crystals were reddish, as was the growing edge of the shell.
The distinctive features of the crystal cell in sectioned
material are the presence of numerous membrane-bound
vacuoles and a complex canaliculus. The vacuoles are few
and small in pre-feeding trochophores (Fig. 8), but they
become larger and more numerous as development pro-
ceeds (Fig. 7). The vacuoles originally contained the crys-
tals that were solubilized during fixation. The appearance
of the vacuolar contents varies with the fixation used, but
they generally appear light under light and electron mi-
croscopy. In material fixed in cacodylate buffered glutar-
aldehyde with ruthenium red and post-fixed in osmium
tetroxide, the contents include small granules, fibrils, and
membranous or myelinic bodies. Larger electron-dense
granules are occasionally found, especially when osmium
tetroxide is used as the primary fixative. These may be
the remnants of partially dissolved crystals.
The canaliculus that also characterizes the crystal cell
is morphologically complex and ends blindly. Its devel-
opment is associated with the development of a ciliated
duct in the pore cell. In the early trochophore, the opening
of the canaliculus faces the pore cell in an area surrounded
by a zonula adherens between these two cells (Fig. 8). At
this stage there is no communication between the lumen
of the canaliculus and the nearby hemocoel. The walls of
the canaliculus possess numerous plications and microvilli
that protrude into the lumen, vastly increasing the surface
area of the canaliculus. Cilia arising from a duct in the
pore cell extend into the lumen of the canaliculus. The
canaliculus can extend for most of the length of the crystal
cell before it ends. It has many short side branches, and
it may bifurcate near its blind end. As development pro-
ceeds, the plications and microvilli become more nu-
310
B R RIVEST
mv
.
*
.
V--
fW*to_
'"PP' w« 5- JBKK'Si/.' ''<\ "-.
i
Figure 6. TEM of the external border of an absorptive cell in a Sear/esia (lira trochophore that had
been soaked in a lerritin solution for 2 h prior to fixation, illustrating the endocytotic uptake of ferritin and
its accumulation in heterophagosomes. Scale bar represents 1 /jm.
Figure 7. TEM showing two crystal cells, a pore ceil and the inner edge of the absorptive cell in a larval
kidney of an early S. dim veliger. Most larval kidney complexes have only one crystal cell. The pore cell at
this stage has little contact with the crystal cell, and has become branched. Scale bar represents 5 jim.
c indicates the canaliculus: cc. crystal cell: ev. endocytotic vesicles; f, ferritin; h. heterophagosomes of
absorptive cell; he. hemocoel; mv, microvilli; pc. pore cell; v, vacuole.
merous and more cilia grow into the canaliculus from the
pore cell as the canaliculus becomes longer (Fig. 9). Oc-
casional invaginations of coated membrane in the cana-
licular wall and nearby coated vesicles suggest receptor-
mediated endocytotic uptake of material from the cana-
licular lumen.
By the time the embryo has developed into an early
veliger, the crystal cell has lost its close association with
the pore cell. These two cells still adjoin each other in a
small area, but junctional complexes between them are
no longer found. With the separation of these cells, the
cilia from the pore cell no longer extend into the canalic-
ulus of the crystal cell. Both the canaliculus of the crystal
cell and ciliated duct of the pore cell now open into the
hemocoel. However, the lumen of the canaliculus has be-
come generally occluded by plications and microvilli.
These structures fill the canalicular opening so that it is
reduced to a web of anastomosing slits (Fig. 10). Only at
one or two sites inside the cell does the lumen of the
canaliculus open up into spaces that are lined by microvilli
(Fig. 1 1 ). At these locations, the contents of the lumen
appear finely granular. Elsewhere there is very little space
within the walls of the canaliculus. Even the finger-like
side branches contain long microvilli that fill the lumen.
A transverse section of these filled branches appears as
two concentric rings of unit membrane (Fig. 1 I ).
Pore cell. The pore cell is the inner most of the three
cells which comprise the larval kidney complex. It is char-
acterized by slit-pores on the cell surface leading to sub-
surface cisternae, an extensive network of membranous
tubules, and one or two ciliated ducts that eventually open
into the hemocoel.
In the prefeeding trochophore, the pore cell is discoid
and adheres closely to the hemocoelic side of the much
larger crystal cell ( Fig. 8 ). It is relatively undifferentiated.
having few if any slit-pores and little tubular network.
At this stage about six cilia arising from a shallow in-
vagination have grown into the canaliculus of the crystal
cell. The invagination deepens and the cilia become
more numerous as the cell continues to differentiate
(Fig. 9). As the crystal cell and pore cell become less
closely associated, the cilia may extend into the hem-
ocoel.
As the pore cell enlarges and differentiates, it grows
away from the crystal cell into the hemocoel and becomes
branched. As seen by Nomarski differential interference
contrast microscopy, the branches attenuate into filopodia
that extend toward the esophagus or to nearby ectodermal
PROSOBRANCH LARVAL KIDNEYS
311
cells. The cell body is thus suspended in the hemocoel
underneath the crystal cell.
The cell surface develops slightly flattened, interdigi-
tating linger-like processes that lie over spaces called sub-
surface cisternae (Skelding and Newell, 1975) (Figs. 12-
14). The subsurface cisternae are typically around 200 to
300 nm deep, but their width varies considerably and in
relation to the number of processes that form their roof.
Although subsurface cisternae eventually develop under
most of the cell surface, they are not interconnected.
The interdigitations are separated by slit-pores of a uni-
form 22.5 to 24.0 nm width that connect the subsurface
cisternae with the hemocoel. The gap may contain fibrous
material, but there was no evidence suggesting that the
slit-pore is spanned by a diaphragm or membrane. An
external lamina covers the surface of the pore cell, al-
though in many preparations it has separated from the
cell surface, possibly a result of fixation (Fig. 14). A dense
material lies beneath the plasmalemma on either side of
the slit-pores (Figs. 12. 13).
The floor of the subsurface cisternae consists of an en-
docytotically active plasmalemma. Imaginations of
coated membrane pinch off to form vesicles that contain
material brought in through the slit-pores (Fig. 14). The
slit-pores could be a site of filtration, although particles
as large as ferritin injected into the hemocoel pass through
the slit-pores into the subsurface cisternae and are taken
up by endocytotic vesicles (Fig. 15). The external lamina,
however, may act as a partial barrier to the passage of
ferritin because the density of ferritin within the subsurface
cisternae was less than that in the hemocoel.
An extensive tubular network is found throughout the
cytoplasm of the pore cell. This anastomosing network
contains material taken in from the subsurface cisternae
by endocytosis. Ferritin injected into the hemocoel is
found within 10 to 15 min in endocytotic vesicles and the
tubular network (Fig. 15). In non-injected embryos, a very
fine fibrillar material is found adjacent to coated
membranes in the subsurface cisternae, in endocytotic ves-
icles, and in some parts of the tubular network (Fig. 16).
Some inflated parts of the tubular network contain dark
bodies. The dark bodies appear homogeneous and consist
of a finely granular material. Small dark bodies may be
surrounded by regions of the lumen of the tubular network
(Fig. 16, near top), but this is not true with large dark
bodies(Figs. 7, 15, 16, near bottom, and 17). Dark bodies
are apparently formed by a condensation of material taken
up endocytotically from the subsurface cisternae and
channeled into the cell by the tubular network. Although
the dark bodies appear electron dense in the transmission
electron microscope, they are colorless and refractile in
live embryos.
Golgi complexes are often found near enlarged regions
of the tubular network (Fig. 17). Small vesicles about 80
nm in diameter frequently found near the Golgi cisternae
may arise from the ends of the Golgi cisternae and fuse
with the tubular network.
During development, the ciliated inpocketing of the
pore cell deepens until the cilia arising from the proximal
end of the duct no longer extend distally beyond the duct
opening into the hemocoel. A second ciliated duct de-
velops, apparently separate from any association with a
canaliculus in the crystal cell. Regardless, it is morpho-
logically indistinguishable from the first duct. As devel-
opment of these ducts proceeds, additional cilia form, both
at the proximal ends and scattered along the walls. Trans-
verse sections of each duct reveal from 10 to 70 cilia. The
number becomes greater in distal regions of the ducts and
with greater cell differentiation. Occasional inpocketings
of coated membrane indicate a low level of endocytotic
activity. The duct walls are surrounded by a continuous
plasmalemma for most of their length, but are perforated
by a few slit-pores in a some distal areas (Fig. 16). In
contrast to the other slit-pores, these are openings between
the duct lumen and the hemocoel. When present. ~ie
are usually only one to five slit-pores in a transverse section
through a ciliated duct. Even when they are more nu-
merous, they have never been seen to occur around more
than one third of the periphery of the duct, nor do they
occur along much of the duct length. Presumably, the
beating of the cilia in the duct could draw hemolymph in
through the slit-pores. However, no ferritin was found in
the ciliated ducts of embryos that had been injected with
ferritin fifteen minutes before fixation.
The internal walls of the duct occasionally possess slit-
pores leading to subsurface cisternae (Fig. 16). These
structures have the same morphology as those on the per-
ipheral cell surface and show some endocylotic activity,
but are not common.
In the mid-veliger stage, prior to the formation of a
mantle cavity into which the cephalopedal elements can
be withdrawn, the distal region of each ciliated duct is
found in a tubular outgrowth from the pore cell. The walls
of this tube may be as thin as 0.3 /jm, but are only oc-
casionally perforated by slit-pores even though some slit-
pores leading to subsurface cisternae are found on the
hemocoelic surface of the tube. As in the more proximal
regions of the duct, some cilia arise within this tubular
section. The ciliated ducts' opening into the hemocoel
has no morphological specializations, although cilia orig-
inating from \\ithin and around the opening extend into
the hemocoel.
The branches of the pore cell containing the ciliated
ducts do not appear to be extending towards a particular
point within the embryo. The distal tubular sections of
the ducts and their openings were found in a variety of
locations: just beneath nearby ectodermal cells, internally
312
B. R. RIVEST
• .*.',., • -**a •'"•'- ••'
. - ••••• . ';«>£. v.*i.
,
.- .
11
"
Figure 8. TEM showing the close association of the developing ciliated duct of the pore cell with the
canaliculus of the crystal cell in an early trochophore ofSearlesia dint
Figure 9. TEM of the same association seen in Figure 8, but in an older trochophore. Numerous cilia
arising from the duct in the pore cell extend into the canaliculus of the crystal cell.
Figure 10. TEM of the opening of the crystal cell's canaliculus into the hemocoel in an early S dim
veliger. The opening of the canaliculus is no longer obstructed by the pore cell, but plications and microvilli
from the canahcular wall reduce the opening to a web of anastomosing slits.
PROSOBRANCH LARVAL KIDNEYS
313
closer to the esophagus, or at some point between the
esophagus and ectoderm.
In living early veligers, the ciliated ducts could be ob-
served with Nomarski differential interference contrast
microscopy. This stage was most suitable for observations
because the larval kidneys were on the transparent neck
region of the embryo. As seen in sections, the branching
pattern and location of the pore cell relative to the crystal
cell varied among embryos and between sides of the same
embryo. Cilia within the ducts were seen actively beating,
but those that extended for most of their length into the
hemocoel beat irregularly and slowly. Occasionally the
beat of the cilia would stop, at which time the ducts could
not be discerned. In some preparations where the embryos
were favorably oriented, two ciliated ducts within a single
pore cell could be distinguished. However, only one could
be discerned in most.
Uptake oj fluorescein-labe lied proteins by the absorptive
cell. The initial endocytotic uptake and concentration of
ferritin by absorptive cells raised questions regarding the
kinds of material that normally would be removed from
the capsular fluid by the these cells. If the larval kidneys
were a storage site for wastes or foreign molecules picked
up from the capsular fluid, then the absorptive cells may
have reacted to the apoferritin protein coat as an exoge-
nous molecule and removed it from the external milieu.
To determine if the absorptive cells take up the endoge-
nous albumen proteins to which they are normally ex-
posed, embryos were soaked in a solution of FITC-labelled
capsular albumen proteins for 10 min to 4 h, then rinsed
with seawater. Under UV epi-illumination, the absorptive
cells on these treated embryos fluoresced, whereas no other
cells did at a visible level (Figs. 18, 19). This was true for
all developmental stages examined, from early prefeeding
trochophores with absorptive cells barely protruding from
the embryo, to late veligers with absorptive cells that
were diminishing in size. No cells in control
embryos examined immediately upon removal from their
capsules fluoresced.
Absorptive cells on embryos soaked in FITC-BSA flu-
oresced in a manner indistinguishable from those exposed
to FITC-conjugated capsule albumen proteins. Therefore,
FITC-BSA was used for many observations on Searlesia
(lira as well as other species. Some embryos, particularly
trochophores during the nurse egg feeding stage of devel-
opment, had swallowed enough of the FITC-BSA solution
so that the gut lumen fluoresced. However, the level of
this fluorescence was comparable to that of the soaking
solution, which was low relative to the brilliance of the
absorptive cells.
The absorptive cells did not take up unconjugated flu-
orescein in quantities visible even after soaking for 24 h.
Furthermore, no fluorescence was seen in the absorptive
cells when glucose, urea, BSA, capsular albumen, or fer-
ritin was added to the fluorescein solution, even when the
ferritin was taken up at levels detectable at the light level.
Contrary to the absorptive cells, the crystal cells did
fluoresce in embryos that had been exposed to fluorescein
solutions. In contrast to FITC-conjugated proteins, flu-
orescein entered the embryos so that the hemocoel flu-
oresced. However, the fluorescence of the crystal cell was
greater than that of the hemocoel. Approximately 50% of
the fluorescing crystal cells possessed an internal structure,
possibly the canaliculus, whose fluorescence was markedly
brighter than the rest of the cell. It was around 5 to 6.5
j/m wide and 30 to 35 nm long, traversing about two-
thirds of the cell. Although this is up to twice the diameter
of the canaliculus visualized in sectioned material, no
other structure in the cell more closely approximates these
dimensions. The vacuoles containing the crystals were
not distinguishable in these treated embryos under UV
epi-illumination.
The pattern of fluorescence in the absorptive cells of
the embryos exposed to FITC-BSA depended on the
length of time of exposure and on the time since they
were removed from the FITC-BSA. After 10 min of ex-
posure, fluorescence could be detected in a few small het-
erophagosomes lying beneath the plasmalemma. With
increasing time of exposure, fluorescence was seen in
many more small heterophagosomes, in larger hetero-
phagosomes further inside the cell and with greater in-
tensity. The larger internal heterophagosomes were dimly
fluorescent within 2 h, and brightly fluorescent after 4 h.
After 7 h, the fluorescence in the larger heterophagosomes
was uneven, suggesting a heterogeneity of protein con-
centrations within those structures. In embryos that were
exposed to FITC-BSA for 3 h and then placed in normal
seawater, few fluorescent small heterophagosomes were
seen after another 3 h, with most of the fluorescence re-
stricted to larger heterophagosomes. The fluorescence
from the FITC-BSA taken in during the 3-h exposure was
still present in the larger heterophagosomes three days
later, the longest embryos survived after excapsulation.
The fluorescence did not appear to diminish in intensity.
Figure 11. TEM of a cross section of a canaliculus in the pore cell of an S. dira early veliger. The lumen
of the canaliculus is surrounded and often occluded by microvilli. There are numerous finger-like side
branches of the canaliculus. each of which is filled by a microvillus.
Scale bars represent 2 /im; ac, absorptive cell; c, canaliculus; cc, crystal cell; cd. ciliated duct; ci, cilia; co.
openings of ciliated duct; en. endoderm; h, heterophagosome of absorptive cell; he. hemocoel; mi, microvilli;
pc, pore cell; sb. side branches of the canaliculus each filled by a microvillus; v, vacuole; za, zonula adherens.
B. R. RIVEST
he
f-
? |
•
_. - /
• • X. >£-^ .i *;/' "T '--,<• ;r*3 •** &.<••>.. .i. t^Vv /
#
\ Mil '-*5t*/ • -~ j
•^isr. '-•' v^-'.&riv.H. ^-^-^H^C&V?/ v
-•: ; is 4^
Figures 12 and 13. TEMs of two sequential grazing sections of the surface of the pore cell in Si'urliviu
dira. illustrating how the interdighations of finger-like processes form uniform 23 nm wide slit-pores and
subsurface cisternae.
Figure 14. TEM of a pore cell in an 5. dira early veliger, showing the extensive tubular network and
endocytotic vesicles forming from the floor of the subsurface cisternae.
Figure 15. TEM of the pore cell of an .f. dira early veliger whose hemocoel had been injected with
ferntin 1 5 min prior to fixation. The ferritin solution still fills the hemocoel. Femtin has entered the subsurface
cisternae and was taken up in endocytotic vesicles which then fused with the tubular network.
Scale bars represent 1 ^m; cd. ciliated duct: db. dark body; e. endocytotic invagination; el. external lamina:
i endocytotic vesicle; f. ferritin; he. hemocoel; m. mitochondrion; pr, cell processes; sc. subsurface cisternae;
sp. slit-pore: tn, tubular network.
nor did it appear in other cells. However, within 1 5 h of
exposure to FITC-BSA, the fluorescence was not clearly
confined to the heterophagosomes; the cytoplasm of the
absorptive cell also fluoresced. although dimly. This may
indicate that labelled material is passed to the cytoplasm,
that some heterophagosomes had burst, or that digestion
of the FITC-BSA was releasing fluorescein which was dif-
fusing into the cytoplasm.
PROSOBRANCH LARVAL KIDNEYS
315
£
Jgfp'"
^-•ffte-'-S •^•iJ/Vfi :**
• f *" <* . > • • • ' A^-'
S.'7 Trf; >N . _L- « ^
MBt.'?-
•^s?, '^-M^^T
^ "^Ri^?«^-;
^isntf
r ^>-/: .- «?"f1r*iX 1
•;*•%
Figure 16. TEM showing the ciliated duct, endocytotic activity and dark bodies in the pore cell of a
Searlesia dira early veliger. Note the early stage of dark body formation within the tubular network near
the upper surface of the cell and the slit-pores within the ciliated duct.
Figure 17. TEM of the pore cell of an early trochophore of S. dira showing a large Golgi complex near
a small dark body (near the upper cell surface) that is surrounded by smaller membrane-bound vesicles of
similarly electron-dense material. Membranes are not always clearly seen around large dark bodies, as is
evident at the bottom of this figure.
Scale bars represent 1 /jm: cd, ciliated duct: db. dark body; e. endocytotic imaginations or vesicles; g,
Golgi complex; he, hemocoel; m, mitochondria; n, nucleus: IT. rough endoplasmic reticulum; sc, grazing
sections of subsurface cisternae: sp. slit-pores; tn, tubular network.
Niicella canaliculata
As in all other species examined, the larval kidney
complexes in A', canaliculala are located laterally in the
neck region. They are first discernible in live embryos
about the time the cephalic vesicle begins to enlarge on
early trochophores. They increase rapidly in size and are
prominent structures on late trochophores (Fig. 20). They
reach their maximum size by the early veliger stage, a
stage by which all nurse eggs have been eaten and the
viscous capsular albumen has been absorbed or eaten.
The A', canaliculata absorptive cells become the largest
cells in the embryo. They are larger than absorptive cells
in any other species examined in this study except for
those in A', lamellosa. which are comparable in size. They
are hemispherical to almost spherical in shape. Their di-
ameter can reach 300 jum, which means the two absorptive
cells can comprise approximately 20% of the tissue volume
of the embryo. The absorptive cells are fragile; their plas-
malemmae are easily ruptured by handling of the embryos
or by osmotic shock, slowly releasing intact heterophago-
somes that may adhere to the outside of the cell (Fig. 2 1 ).
Only a general ultrastructural examination of the N.
canaliculata larval kidney was completed because a sat-
isfactory fixative was not found. Early trochophores
through late veliger stages were sectioned for examination
with light microscopy, but only the late trochophore stage
was serially thin sectioned for TEM. The larval kidney in
jV. canaliculata consists of the same arrangement of an
absorptive cell, a crystal cell and a pore cell as found in
S. (lira. The cytology of the absorptive cell in Ar. canali-
culata is also similar. The external plasmalemma with its
microvilli (Fig. 21) is endocytotically active. The heter-
ophagosomes are heterogeneous in their density at both
the light (Fig. 22) and TEM level. Rosettes of glycogen
are present in the cytoplasm. Ferritin is rapidly taken in
and stored in heterophagosomes, as is FITC-BSA (Fig.
23). These cells are endocytotically active throughout the
intracapsular veliger phase of development. The crystal
and pore cells differentiate in association with each
316
B. R. RIVEST
ec— •
Figure 18. Light photomicrograph of a live Searlesia dim early veliger that had been soaked in a solution
of FITC-capsular albumen for 3 h.
Figure 19. Epifluorescence photomicrograph of the same veliger seen in Figure 18, showing the bright
absorptive cells that accumulated the fluorescein-labelled protein.
Scale bar represents 100 ^m; ac. absorptive cell; ec, ectoderm; f, foot; ne, swallowed nurse eggs; v, velum.
other and develop the same ultrastructures they do in
5. dim.
Presence ofahsor/nive cells in di/fercnt gastropod spe-
cies. Embryos and some larvae from the three gastropod
subclasses were tested for the presence of larval kidney
absorptive cells. The species examined normally hatch
either as veligers or postmetamorphic juveniles. Embryos
were removed from their capsules, rinsed in seawater and
placed in an FITC-BSA solution for at least I h before
being examined under UV epi-illumination. Posthatching
veligers of several species were obtained from cultures for
testing.
In all seven prosobranch species tested that hatch as
veligers, paired absorptive cells were found in prehatching
stages (Table I). After hatching, veligers lost their absorp-
tive cells within two or three days. Crystal cells were also
detected in Oenopota levidensis veligers with polarized
light. These cells were no longer detectable after the ab-
sorptive cells were resorbed during the second to third
day after hatching.
Of the twelve prosobranch species tested that hatch as
juveniles, only ten had embryos possessing paired ab-
sorptive cells (Table I). In these species, the absorptive
cells were resorbed or at least did not protrude from the
embryos after intracapsular metamorphosis. Littorinu
sitkanu and Petaloconchus moniereyensis embryos were
the only prosobranchs that did not show absorptive cells
at some stage. The capsules of L. silkana contain albumen,
but those of P. nn >ntereyensis possess little or none as they
are packed with nurse eggs. None of the opisthobranch
or pulmonate embryos had absorptive cells.
Discussion
Lack of evidence of waste accumulation
Marine prosobranchs are basically ammoniotelic, al-
though uric acid may be found in various tissues (Nicol,
1960; Duerr, 1968). The larval kidneys, however, do not
contain uric acid. The histochemical tests for urates in
Searlesia dim and Nncella canaliculata embryos were
negative, and there was no ultrastructural evidence for
urates like that seen in terrestrial snails by Bouillon and
Vandermeerssche (1962) and Pecheco (1971). The lack
of uric acid and the positive PAS staining of the absorptive
cells indicates that their large size is due to polysaccharides
or polysaccharide-protein complexes and is unlikely due
to accumulation of waste products.
Although no evidence for excretory activity by the larval
kidneys was found during this study, previous reports in-
dicating such a function may have been a consequence
of several factors. First, these complexes of cells were called
larval kidneys without any apparent definitive elucidation
of their function, thus possibly biasing others as to their
supposed role. Second, the delicate nature of the absorp-
tive cells and the difficulty in fixing them well for ultra-
structural studies help explain some of the observations
of 'excretory' activity that have been reported. These large
cells often break open, releasing intact heterophagosomes,
before other structures visibly deteriorate in embryos re-
moved from their capsules. Thus, the "excretory granules"
seen by D'Asaro (1966, p. 895) emanating from the larval
kidneys of Thais haeimistoma and what Portmann (1930)
thought were escaping waste-ladened wandering cells in
Biieeinum iindatitin embryos were likely to be hetero-
phagosomes spilling from ruptured absorptive cells. The
fragility of these cells may be due to the proteins they
have accumulated causing osmotic shock when the em-
bryos are removed from their capsules. This may be why
these cells are difficult to fix well for ultrastructural study.
The exocytosis cited as evidence for excretion by the ab-
sorptive cells in Nncella lapillns (Fioroni, 1985; Fioroni
el u/-. 1985) is similar to that seen in poorly fixed embryos
PROSOBRANCH LARVAL KIDNEYS
317
Figure 20. SEM of a A'/<a'//a canaliculata late trochophore showing the large size of the absorptive cells.
Figure 21. SEM of an absorptive cell on an early veliger of A'- canaliailuia. The cell has broken open
revealing heterophagosomes. Microvilli are evident covering the external surface of the cell.
Figure 22. A light photomicrograph of a l-/tm thick section of an absorptive cell in an early veliger of
N. canaliculata. Although the heterophagosomes are evident, the internal ultrastructure of the cell was
greatly affected by fixation artifacts.
Figure 23. An epifluorescence photomicrograph of an absorptive cell on a live A', canaliculata early
veliger that had been soaked for 1 h in a solution containing FITC-BSA. Heterophagosomes are clearly
visible, with their level of fluorescence differing according to the amount of FITC-BSA they contain.
Scale bars represent 100 /jm; ac. absorptive cell: cv. cephalic vesicle; ec. ectoderm: en. endoderm; f, foot;
h. heterophagosome; sh, shell: v. velum.
of S. (lira and A', canaliculata. It would seem illogical to
describe the larval kidneys as specialized organs for en-
capsulated development that accumulate waste and then
release it into the capsular fluid that bathes the embryos.
Absorptive cell
The absorptive cell is the only cell of the larval kidney
complex that is exposed to the capsular fluid. It is clear
from the experimental exposure of S. (lira embryos to
ferritin and FITC-capsular albumen that the absorptive
cells rapidly take up protein from the external milieu by
endocytosis and store it in heterophagosomes.
The early differentiation of the absorptive cells is con-
sistent with the hypothesis that the albumen is an impor-
tant source of nutrition for which capsulemates compete.
The earlier an embryo has functional absorptive cells the
318
B. R. RIVEST
Table I
The presence of paired, laterally located fluorescent ectodermal absorptive cells in gastropod embryos or larvae exposed to FITC-BSA
Species (hatching stage*)
Fluorescing cells
Developmental stages tested
PROSOBRANCH1A
Mesogastropoda
Lacuna variegala (V)
Lil/orina silkana (J)
Pelaloconchus montereyensis (J)
Calyptraea fastigiata (V)
Crepidula fonncala (V)
Crepidula adunca (J)
Trichoimpis camvllaia ( V )
Lamellaria sp. (V)
Neogastropoda
Nucclla canaliculate (J)
NHcellii cnuirginala (J)
Nucella lamellosa (J)
Nucclla lapillus (J)
Nucella lima (J)
Ocenebra japonica (J)
Ceralostnma foliatuin ( J )
Amphissa columbiana (V)
Scarlcsia dira (J)
Neptunea lyrala (J)
Oenopota Icvidensis (V)
OPISTHOBRANCH1A
Onchidoris bilamellata (V)
Tn Ionia diomedea (V)
PULMONATA
Lymnaea stagnalis (J)
+ (weak)
+ (weak)
Prehatching and 2-day-old veligers
Veligers
Before, during, and just after feeding on nurse eggs
Late trochophore
Prehatching veligers
Early to late shell formation
Prehatching to 3-day-old veligers
Prehatching veligers
Trochophores to late veligers
Trochophores, late veligers
Trochophores to late veligers
Trochophores to early veligers
Trochophores to early veligers
Early veligers
Early trochophores
Prehatching veligers
4 days post-hatching
Trochophores to late veligers
Late trochophores
0-2-day-old veligers
Veligers older than 2 days
1-2 days posthatching
Prehatching veligers
Before and during intracapsular albumen ingeslion phase
V = veliger; J = crawling juvenile.
greater are its chances of getting a larger share of the al-
bumen in the capsule. The larval kidneys may be present
as early as before (Franc, 1940) or just after (Eisawy and
Serial, 1974) the end of gastrulation. In living 5. dira and
N. canaliculata embryos, the absorptive cells are visible
with light microscopy and are increasing in volume by
the early trochophore stage. The cells reach their maxi-
mum size by the early veliger stage, a time by which no
viscous albumen remains in the capsular fluid. There are
great differences in absorptive cell sizes among proso-
branch species, but this does not necessarily mean differ-
ences in the amount of albumen available. In species with
smaller absorptive cells, the absorbed proteins may be
passed to the rest of the embryo more rapidly or sooner
than in species with larger absorptive cells.
It appears that when the larval kidneys of 5. dira and
N. canaliculata diminish in size they are totally resorbed,
but the cytology of post-resorption stages has not been
examined. No ultrastructural evidence of lysosomal ac-
tivity was found in S. dira absorptive cells up through
mid-veliger, the oldest stage examined. Resorption occurs
during intracapsular metamorphosis in all the neogastro-
pod species examined in this study that hatch as juvenile
snails. However. Conklin (1897, p. 143) said the larval
kidneys in Crepidula ionttcata embryos "appear to be
pinched off completely" and in C'. adnnca, Moritz ( 1939)
said they were rolled off by the advancing mantle edge on
the left side and by the larval heart on the right. In my
examinations of these two species and on the confamilial
Calyptraea fastigiata, the larval kidneys became almost
spherical and then were easily dislodged when the capsules
were opened. Observations through the clear walls of
carefully handled capsules never revealed cast off absorp-
tive cells. Rather, they were resorbed.
For species that hatch as veligers, the time of absorptive
cell resorption varies. In Thais savignyi (Eisawy and Sorial,
1974) and T. hacmasionui (Belisle and Byrd, 1980), they
are resorbed before hatching. If the primary function of
the absorptive cell is the uptake of capsular albumen dur-
ing encapsulated development, then they are no longer
needed after hatching. However, some species retain their
absorptive cells for a few days after hatching. Tests with
FITC-BSA indicate that the veligers of Lacuna vancgata,
Trichotropis canccllata. and Oenopota Icvidcnsis have
PROSOBRANCH LARVAL KIDNEYS
319
functional absorptive cells for at least two or three days
after hatching. Similarly, the larval kidneys of Trivia eu-
ropea persist after hatching, but only for a short time (Le-
bour, 193 1 ). Thus, the absorptive cells appear to be prin-
cipally an adaptation for intracapsular development.
The tests using FITC-BSA on a few opisthobranch and
pulmonate species did not reveal structures that absorbed
and concentrated proteins as did the prosobranch ab-
sorptive cells. Opisthobranch veligers may have an un-
paired ectodermal larval or secondary kidney that consists
of several types of cells near the anus (Bonar and Hadfield,
1974; Bickell and Chia, 1979). The veligers of Aeolidia
papillosa are reported to possess protonephridia (Barto-
lomaeus, 1989). In pulmonates, some albumen is taken
up generally during early embryogenesis by endocytosis,
but most of the albumen is ingested after the gut becomes
functional (Raven, 1946, 1975). Thus, prosobranch ab-
sorptive cells, as specialized embryonic ectodermal struc-
tures for the uptake of capsular albumen, may be unique
among the invertebrate phyla.
Crystal cell
Although the function of the absorptive cell appears to
be the uptake and storage of capsular albumen, the func-
tions of the crystal and pore cells are less clear. The as-
sociation of the crystal cell with most of the absorptive
cell's internal surface suggests that the functions of these
cells may be related, but evidence for this is presently
lacking. The association may simply reveal an ancestral
relationship.
In the crystal cell, vacuoles containing the crystals in-
crease in number as the embryos develop. The crystals
are likely to be made of a calcium salt. They are possibly
calcium carbonate, for they were more soluble in acids
than bases, were quickly dissolved in a solution of EGTA,
which chelates Ca++ ions (Schmid and Reilley, 1957), and
were stained by alizarin red S, which stains bone (Emmel
and Cowdry, 1964) and calcium carbonate (Buddemeir
and Kinzie, 1976). The crystals are in a readily soluble
form, as indicated by their rapid dissolution after vacuolar
membrane breakdown in non-salt-saturated solutions.
The vacuolar membrane clearly protects the integrity of
the crystals. The EGTA solutions did not affect the crystals
in live embryos, but quickly dissolved those in heat-killed
embryos. Although the crystals are more likely to contain
calcium than uric acid, which dissolve more readily in
high pH solutions (Thorpe. 1930), the function of se-
questering calcium is unclear.
Endocytotic activity along the canaliculus in the crystal
cell suggests that the contents of the vacuoles may origi-
nate from the canalicular lumen. However, the evidence
indicates a very low level of endocytotic activity, and vac-
uoles are formed both before and after the separation of
the pore cell and crystal cell when the canaliculus is first
open to the hemocoel. Thus the relationship between the
function of the canaliculus and crystal formation is un-
certain. The canaliculi in the parietal cells of the vertebrate
gastric mucosa are implicated in ion transport for HC1
secretion (Ito and Winchester. 1963). However, the can-
aliculi in the crystal cells of S. dira are unlikely to be
heavily involved in ion transport, for they lack the close
association with numerous mitochondria that is charac-
teristic in cells that do (Lawn, I960; Copeland, 1967; Sar-
det et al.. 1979). Although the large surface area of the
canaliculus due to the numerous microvilli and plications
suggests a membrane-dependent function, the restricted
nature of the lumen indicates an impeded movement of
material either into or out of the canaliculus.
In experiments exposing 5. dira embryos to solutions
of tluorescein. the crystal cells fluoresced brightly. How
the fluorescein is concentrated and what role the canaliculi
play in the concentrating mechanism remains unknown.
Crystal cells have been reported in several other pro-
sobranch embryos (Portmann, 1930; Franc, 1940;Fioroni,
1966, 1985). Franc ( 1940) described a vacuolated cell lying
under the large external larval kidney cell in the embryos
of Oceanebra aciculata. As in S. dira, the cell increased
in size and became filled with vacuoles, but the vacuoles
in O. aciculata became greenish in older developmental
stages. In most of the species listed in Table I, the larval
kidney cells had a yellowish color with the crystals in the
crystal cells being a dark yellow to yellow-green. However,
in additional observations on embryos of Oceanebra (ei-
ther O. interfossa or O. lurida), the crystal cells were
greenish-yellow with green granules (pers. obs.). There was
no green color in young embryos, nor was there any green
color elsewhere in the capsule. Therefore, the pigment
likely was synthesized dc novo.
Pore cell
Pore cells are a conspicuous cell type in the connective
tissue of gastropods and bivalves (Sminia and Boer, 1973).
There are also pore cells, referred to as nephrocytes or
podocytes, in the blood spaces of decapod gill shafts
(Wright, 1 964; Strangways-Dixon and Smith, 1970; Foster
andHowse, 1978: Taylor and Greenway, 1979). 'Nephro-
cyte' is an appropriate label for these cells if the material
they remove from the hemolymph is an excretory product,
but pore cells have also been implicated in the production
of blood pigments in mollusks (Sminia and Boer, 1973;
Skelding and Newell. 1975). Calling them 'pore cells" aptly
describes their distinctive morphological features without
implying function, whereas calling them 'podocytes' is
misleading because of their morphological and functional
differences with vertebrate kidney podocytes (Bloom and
Fawcett. 1968).
320
B. R. RIVEST
In 5. dim pore cells, material from the hemolymph
passes through the slit-pores and is taken into the pore
cell by endocytotic vesicles that form from the basal
membranes of the subsurface cisternae. The external
lamina may act as a filter, for in experimentally injected
5. dira embryos the density of ferritin molecules inside
the subsurface cisternae was markedly less than that in
the hemocoel. Additional filtration is unlikely at the slit-
pores, because their gaps are much wider than the di-
ameter of the ferritin molecules. Furthermore, a mem-
brane or diaphragm like that found in the nephrocytes of
shrimp gills (Foster and Howse, 1978) was not seen span-
ning the slit-pores in S. dira. The material taken in by
endocytosis is transferred to an extensive network of tu-
bules, where it appears to be condensed to form large
dark-staining bodies. A similar ultrastructure is found in
pore cells in the gill blood vessels of a land crab, Hol-
thitisana tmnsversa (Taylor and Greenway, 1979). In
general, pore cells appear to remove material from the
hemolymph. However, the nature of this material and its
fate remains to be elucidated.
A characteristic of the pore cells of S. dira and N. can-
aliculata that has not been found in pore cells in other
species is the ciliated ducts. In S. dira, only the first of
the two ciliated ducts develops in association with the
canaliculus of the crystal cell, but eventually the two ducts
are indistinguishable. They both originate near the cell
center, contain a similar number of cilia, possess some
slit-pores leading to the hemocoel somewhere along their
length, and open into the hemocoel. The duct lumens in
most places are only slightly wider than the cilia they con-
tain. Movement of the cilia could move fluid down and
out of the ducts. However, the low number of slit-pores
to the hemocoel in the duct walls would result in a low
flow rate. Indeed, no ferritin was found in ducts in injected
embryos.
Comparison with protonephridia
It is logical to compare the 'larval kidneys' of S. dira
and N. canalicidata to protonephridia because of their
name, their bilateral location, and the fact that protone-
phridia have been described in other prosobranch, opis-
thobranch, and pulmonate gastropods, in chitons, and in
bivalves (Erlanger, 1894; Brandenburg, 1966; Bartolo-
maeus, 1989; Ruthensteiner and Schaefer, 1991). This
comparison is especially attractive in light of the pore
cells' ciliated ducts with their slit-pore perforations to the
hemocoel being so similar in morphology to many pro-
tonephridia. Although protonephridia are diverse in de-
tails of their morphology (Goodrich, 1945; Brandenburg.
1966, 1975; Wessing and Polenz, 1974; Wilson and
Webster, 1974; Ruppert and Smith. 1988), they basically
consist of a terminal cell or cells with one or more cilia
within a cytoplasmic tube that leads to a channel cell or
tubule that opens externally. The beating of the cilia draws
hemolymph into the tubule through a weir of slits formed
in some species by the interdigitation of microvilli from
both the terminal and channel cells (Brandenburg, 1975;
Kummel, 1975).
It is clear that the larval kidneys of 5. dira and N. can-
alicidata are not functional protonephridia because a fil-
tration weir never develops and the ciliated ducts lead not
to the external milieu, but to the haemocoel. However, it
is possible that the larval kidneys may have evolved from
protonephridia and that the components have become
disorganized and functionally altered. If this is so, the
pore cell probably evolved from a protonephridial ter-
minal cell. The pore cell's ciliated ducts are reminiscent
of protonephridial 'flames.' and the slit-pores and sub-
surface cisternae are similar to that seen in a section of
the terminal cell of the trochophore of Pomatoceros tri-
cjiteter (Wessing and Polenz, 1974, Fig. 2b). The crystal
and absorptive cells may have evolved from channel cells.
The canaliculus may be the last remnant of the crystal
cell's ancestral duct, with no such vestige remaining in
the highly altered absorptive cell.
The presence of larval protonephridia may represent
the ancestral condition in mollusks (Bartolomaeus, 1989),
but gastropod embryos undergoing encapsulated devel-
opment experience an environment different from the
open-ocean surroundings experienced during develop-
ment of their free-spawning, more primitive relatives. The
egg capsules with their intracapsular fluid presumably
create an osmotic environment different from that of sea-
water. This is supported by two observations made during
this study. One is the fixation artifacts seen. These artifacts
suggested an osmotic problem and were diminished when
fixatives with osmolarities higher than seawater were used.
The second is the bursting of the absorptive cells when
embryos were carefully excapsulated into seawater. This
was most noticeable in species where these cells are par-
ticularly large, and is presumably due to the osmotic im-
balance between the concentrated proteins within the
heterophagosomes and the seawater. Because protone-
phridia are reported to be involved in osmoregulation
(Braun cl al., 1969) and the filtration and reabsorption
of macromolecules from coelomic fluid (Smith and Rup-
pert, 1988). it would seem logical for encapsulation to
reduce the selective pressures that maintain protone-
phridia as osmoregulatory organs. It is thus possible that
the larval kidneys described herein may represent rudi-
mentary protonephridia. The lack of organization among
the components is not surprising if the need for functional
protonephridia is lacking. The vestigial eyes of cave fish
show a similar disorganization of components (Remane,
1971). The functions of the larval kidneys' cells also have
changed. The pore and crystal cells have functions that
PROSOBRANCH LARVAL KIDNEYS
321
are presently unknown, and may simply be vestigial
structures. The absorptive cell, however, has taken on a
new role. It does not accumulate wastes as previously
thought (('.#.. Franc, 1940). but is specialized for the ac-
quisition of capsular albumen.
An ultrastructural study of the ontogeny of the larval
kidney complex in Nassarius reticitlatux might shed light
on this question of homology, for Ruthensteiner and
Schaefer (1991) found in their light microscopical study
protonephridia with what appears to be an enlarged ab-
sorptive cell adjacent to the excretory pore. A similar study
on the early ontogeny of the larval kidney complex of 5.
dira or N. canaliculata might help reveal the germ layer
origins of the component cells and their homology with
cells in other protonephridia. Investigators have described
the larval kidneys as being ectodermal or also part me-
sodermal (Hey nions. 1893;Conklin, 1897; Casteel, 1904;
Pelseneer, 1911; D'Asaro, 1966). The absorptive cell in
5. dira embryos is probably derived from an ectodermal
cell for it has junctional complexes with the surrounding
unspecialized ectodermal cells. However, the origins of
the inner two cells is unclear. If they are protonephridial
in nature, then they may be ectodermal (Goodrich, 1945).
However, the pore cells in 5. dira larval kidneys have
similarities with pore cells in the connective tissues of
other gastropods and bivalves (Sminia and Boer, 1973),
which suggests that they may be mesodermal in origin. It
should be noted here that protonephridial terminal cells
(presumably ectodermal in origin) in the larvae of the
polychaete Sabellaria transform into a podocyte (pre-
sumably a mesodermal cell) in the coelomic lining at
metamorphosis (Smith and Ruppert, 1988), indicating
that caution must be exercised in making assumptions
about germ layer affinities (Ruppert and Smith. 1988).
Nutritional role of capsular albumen and other
suggested functions
Although albumen in prosobranch egg capsules does
not appear to be bactericidal (Rivest, 1981; Pechenik ct
ai. 1984), it may serve several other functions. It may be
important physically by increasing the viscosity of the in-
tracapsular fluid relative to seawater so that embryos can
maneuver and teed on nurse eggs (Rivest, 1983). It may
be important osmotically by protecting embryos from os-
motic shock (Pechenik, 1983) or by aiding hatching (Her-
tling, 1928). The albumen, containing proteins and car-
bohydrates and sometimes lipid (Bayne. 1968; Stock-
mann-Bosbach and Althoff. 1989), also can have nutritive
value for prosobranch embryos. It can influence the de-
velopmental stage that hatches (Giglioli, 1955) or the size
at hatching (Rasmussen. 1951; Rivest. 1986). Its nutri-
tional value is suggested also by its endocytotic uptake in
early cleavage stages (Elbers and Bluemink, 1960; Fioroni,
1977) and by the presence of a special transitory albumen
digestive sac in the embryonic gut of some species (Port-
mann, 1955; Portmann and Sandmeier, 1965). In the
closely allied pulmonates, protein gains in the growing
embryo are positively correlated with protein losses in the
capsular fluid (Morrill, 1964). Thus, it is apparent that
regardless of its other functions, the proteinaceous albu-
men can be important nutritionally.
The albumen taken up by the absorptive cells in pro-
sobranch larval kidneys is presumably used for embryonic
nutrition, but the nutritional significance of this absorbed
albumen is presently unknown. Assuming that the abor-
tive cells' volume is due primarily to absorbed albumen
proteins, the fact that these cells in N. canaliculata can
comprise 20% of an embryo's tissue volume suggests that
in this species the larval kidneys may procure a significant
quantity of albumen. It would be difficult to empirically
examine this in Nucella unless techniques for culturing
encapsulated embryos improve. Although early embryos
of Conns pennaceus can be reared to the veliger stage
outside their capsules with no apparent detrimental effect
(Perron, 1981), young Nucella embryos do not survive
very long once removed from their capsule (Pechenik el
ai, 1984; Stockmann-Bosbach, 1988; pers. obs.).
Acknowledgments
This study benefitted from the advice of R. R. Strath-
mann. the late R. L. Fernald, the late C. G. Reed, C.
Gabel, E. M. Eddy, and R. A. Cloney, who piqued my
curiosity about prosobranch larval kidneys. E. Ruppert
made numerous constructive criticisms on the manu-
script, especially regarding the possible vestigial nature of
the larval kidneys. R. Shimek provided Oenopota levi-
c/ensis veligers. and R. Palmer provided Nucella lima egg
capsules. The director of the Friday Harbor Laboratories
kindly provided use of the laboratories' facilities. This
study was supported in part by National Institutes of
Health grants 5-T01-HD00266-10 and 1-T32-HD07183-
0 1 and National Science Foundation grant OCE 78 1 8608
(to R. R. Strathmann).
Literature Cited
Bartolomaeus, T. 1 989. Larvale Nierenorgane bei Lepidxhiton cinemts
(Polyplacophora) und Acoliilia papillosa (Gastropoda). Zoonior-
phology 108: 297-307.
Bayne, C. J. 1968. Histochemical studies on the egg capsules of eight
gastropod molluscs. Proc. Malacol. Sac. Loud. 38: 199-212.
Belisle, B. VV., and \V. Byrd. 1980. In vitro egg activation and matu-
ration, and ultrastructural analysis of development in the marine
prosobranch. Thais haemastoma Trans. Am .Microsc. Soc. 99: 111-
127.
Bickcll, L. R., and K. S. Chia. 1979. Organogenesis and histogenesis
in the planktotrophic veliger of DoriMla xtcinhcrgae (Opisthobran-
chia: Nudibranchia). Mar. Binl. 52: 291-313.
322
B. R. RIVEST
Bloom, W., and D. W. Fawcett. 1968. A Textbook of Histology. 9th
cd. W. B. Saundt-rs Co.. Philadelphia.
Bobretzky, N. 1877. Studien iiber die embryonal Entwicklung der Gas-
tropoden. Arch. Mikrobioi Anal. Bd. 13: 95-169.
Bonar, I). B., and M. G. Hadfield. 1974. Metamorphosis of the marine
gastropod Phestilla sibogae Bergh (Nudibranchia: Aeolidacea) I. Light
and electron microscopic analysis of larval and metamorphic stages.
./. E\p. Mar Biol. Ecol 16: 227-255.
Bouillon, J., and G. Vandermeerssche. 1962. Electron microscope ob-
servations of the excretory organ of the pulmonary molluscs. 5th Int
Cong. Electron Microsc.. Philadelphia 2: WW14-15.
Brandenburg, J. 1966. Die Reusenformem derCyrtocyten. Zoo/. Beit.
12: 346-417.
Brandenburg, J. 1975. The morphology of the protonephndia. Fortsch.
Zool. 23: 1-17.
Braun, G., G. Kummel. and J. A. Mangos. 1966. Studies on the ultra-
structure and function of a primitive excretory organ, the protone-
phridium of the rotifer Asplanchna pnodonta. Pflugers Arch 289:
141-154.
Buddemeir, R. \\ ., and R. A. Kin/.ie, III. 1976. Coral growth. Oceanogr.
Mar Biol Ann Rev 14: 183-225.
Casteel, D. B. 1904. The cell-lineage and early larval development of
Fiona manna, a nudibranch mollusk. Proc Acad. Nail. Sci. Phil.
Set: 3 56: 325-405.
Cavey, M., and R. Cloney. 1972. Fine structure and differentiation of
ascidian muscle. I. Differentiated caudal musculature of Dislaplia
occidentali.i tadpoles. J Morphol. 138: 349-374.
Conklin, E. G. 1897. The embryology of Crepidula. J. Morphol 13:
1-226.
Copeland, D. K. 1967. A study of salt secreting cells in the brine shrimp
(Anemia sa/ina). Proloplasma 63: 363-384.
D'Asaro, C. N. 1966. The egg capsules, embryogenesis. and early or-
ganogenesis of a common oyster predator, Thais haemastoma (Gas-
tropoda: Prosobranchia). Bull. Mar Sci 16: 884-914.
Duerr, F. G. 1968. Excretion of ammonia and urea in seven species
of marine prosobranch snails. Comp Biochcm. Phr.siol 26: 1051-
1059.
Eisawy, A. M., and A. E. Sorial. 1974. Egg capsules and development
of Thais savignyi Deshayes. Bull Insl Oceanogr. Fish.. Cairo 4:
237-258.
Elbers, P. F., and J. G. Bluemink. 1960. Pinocytosis in the developing
egg of Limnaca siagnc/is L. Exp Cell Res. 21: 619-622.
Kmmel, V. M., and E. V. Cowdry. 1964. Laboratory Technique in Bi-
ology and Medicine The Williams and Wilkens Co.. Baltimore.
Erlanger, R. 1894. Zur Bildung des Mesoderms bei der Pahidnia vi-
\-ipara Morph Jb 22: 113-118.
Fioroni, P. 1966. Zur Morphologie und Embryogenese des Darmtraktes
und der transitorischen Organe bei Prosobranchiern (Mollusca, Gas-
tropoda). Rev. Suisse Zool. 73: 621-876.
Fioroni. P. 1977. On the penphenc resorption of albumen in gastropods
and its significance for general embryology. Zool Jb Anal. Bd 98:
181-202.
Fioroni, P. 1985. Struktur und Funktion der larvalen Zellen des Ceph-
alopodiums bei jungen intrakapsularen Larven von Nucella lapillus
(Gastropoda. Prosobranchia). Zoo/ Bcitr N. F 29: 103-1 17.
Fioroni, P., G. Sundermann, and D. P. Scheidegger. 1984. Die Ultra-
strukturder freien Rhogocyten bei intrakapsularen Veligern von Nu-
ce/la lapillus (Gastropoda, Prosobranchia, Stenoglossa). /.ool An:
212: 193-202.
Fioroni, P., D. P. Scheidegger, and G. Sundermann. 1985. Die Ultra-
struktur der Larvalnierne bei intrakapsularen Larven von Nucella
lapillus (Gastropoda, Prosobranchia, Stenoglossa). Zool Jb dial
113: 145-164.
Foster, C. A., and H. D. Howse. 1978. A morphological study on gills
of the brown shrimp. Penaeus a:/ecus. Tissue Cell. 10: 77-92.
Franc, A. 1940. Recherches sur le developpement A'Ocmebra aciculata
Lamark. Bull. Biol. France Belgiaue 74: 327-345.
Franc, A. 1941. Sur les reins larvaires de certains Mollusques proso-
branches. C. R Soc Biol. Paris 135: 1487-1489.
Giglioli, M. E. 1955. The egg masses of the Naticidae (Gastropoda).
J. Fish. Res Board Can. 12: 287-327.
Glaser, O. 1904. Excretory activities in the nuclei of gastropod embryos.
Am Nat. 38: 513-516.
Goodrich, E. S. 1945. The study of nephridia and genital ducts since
1895. O. J. Microsc. Sci. 86: 113-392.
llertling, H. 1928. Beobactungen und Versuche an den Eiern von Lit-
torina und Lacuna. Bedeutung der Eihiillen. Entwicklung im natur-
lichen und abgeanderten Medium, tt'iss. Meers. Abl. Helgoland 17:
1-49.
Heymons, R. 1893. Zur Entwicklungsgeschicte von Umbrella medi-
lerranea. Lam. 7. H'iss. Zool. 56: 245-298.
Humason, G. L. 1972. Animal Tissue Technique. 3rd ed. W. H. Free-
man and Co.. San Francisco.
Ito, S., and R. J. Winchester. 1963. The fine structure of the gastric
mucosa in the bat. J. Cell Biol 16: 541-577.
Kummel, G. 1975. The physiology of protonephndia. Forlsch. Zool.
23: 18-32.
Lawn, A. M. 1960. Observations on the fine structure of the gastric
parietal cells of the rat. J. Biophys. Biochem Cylol 7: 161-166.
Lebour, M. 1931. The larval stages of Trivia europea. J. Mar. Biol.
Assoc. L'. A 17: 819-831.
Luft, J. H. 1961. Improvements in epoxy resin embedding methods.
J. Biophys Biochem Cylol 9: 409-414.
Moritz, C. E. 1939. Organogenesis in the gasteropod Crepidula adunca
Sowerby. L'mv Calf- Publ. Zool. 43: 217-248.
Morrill, J. B. 1964. Protein content and dipeptidase activity of normal
and cobalt-treated embryos of Lymnaea palustris. Acta Embryol.
Morphol Exp 1: 131-142.
Nicol, J. A. C. 1960. The Biology of Marine Animals. Interscience
Publishers, New York.
Pearse, A. G. E. 1972. Hisiochemislry. I'ol. II. 3rd ed. The Williams
and Wilkens Co., Baltimore.
Pecheco, J. 1971. Fine structure of concretions in the kidney epithelium
of Pomacea urccus (Gastropoda: Prosobranchia). Acta Biol I'ene:.
7: 443-458.
Pechenik, J. 1983. Egg capsules of Nucella lapillus (L.) protect against
low-salinity stress. J. Exp. Mar. Biol. Ecol. 71: 165-179.
Pechenik, J. A., S. C. Chang, and A. Lord. 1984. Encapsulated devel-
opment of the marine prosobranch gastropod Nucella lapillus. Mar.
Biol 78: 223-229.
Pelseneer, R. 1911. Recherches sur I'embrvologie des gastropodes. ,-lcarf.
R Belg. Mem Cl. Sci. Collect 3: 1-167.
Perron, F. E. 1981. Larval growth and metamorphosis of Comix (Gas-
tropoda: Toxoglossa). Pac Sci- 35: 25-38.
Portmann, A. 1930. Die Larvenniern von Biiccimim undatum L. Z.
Zell Mikro Anal 10: 401-410.
Portmann, A. 1955. La metamorphose "abritee" de Fnsus (Gast. Pro-
sobranches). Re\'. Suisse Zool. 62: 236-252.
Portmann, A., and E. Sandmeier. 1965. Die Entwicklung von Vorder-
darm, Macromeren und Enddarm unterdem Einflussvon Nahreiern
bei Biiccinuni. Murcx und Nucella (Gastrop. Prosobranchia). Rev.
Suisse Zool 72: 187-204.
Rasmussen, E. 1951. Faumstic and biological notes on marine inver-
tebrates II. The eggs and larvae of some Danish Marine Gastropods.
I'ulcnsk. Mcdd Dan Nalurhist Foren 113: 201-249.
Raven, C. P. 1946. The development of the egg of Lymnaea stagnalix
L. from the first cleavage till the trochophore stage, with special ref-
erence to its "chemical embryology." Arch. Neerl. Zool. 7: 353-434.
PROSOBRANCH LARVAL KIDNEYS
323
Raven, C. P. 1975. Development. Pp. 367-400 in Pulmonates. V.
Fretter and J. Peake, eds. Academic Press, New York.
Remanc, A. 1971. Die (Jnmdlagen des natiirlichen Systems, der ver-
gleiehemleii Anatomic und der Phylogenetik. Otto Koeltz, Koenig-
stein-Taunus (reprint of the 1956 ed. published by Geest & Portig.
Leipzig).
Reynolds, K. S. 1963. The use of lead citrate at high pH as an electron-
opaque stain in electron microscopy. J. Cell Biol 17: 208-21 1.
Richardson, K. C., L. Jarrett, and E. H. Kinkc. 1960. Embedding in
epoxy resins for ultrathin sectioning in electron microscopy. Slain
Techno/. 35: 313-323.
Rivest, B. R. 1981. Nurse egg consumption and the uptake of albumen
in the embryonic nutrition of marine snails. Ph.D. dissertation. Uni-
versity of Washington. Seattle.
Rivest, B. R. 1983. Development and the influence of nurse egg allot-
ment on hatching size in Searlesia dim (Reeve, 1 846) (Prosobranchia:
Buccinidae). J. Exp. Mar Biol. Ecol. 69: 217-241.
Rivest, B. R. 1986. Extra-embryonic nutrition in the prosobranch
gastropod i'rosalpin.\ cinerea (Say. 1822). Bull. Mar. Sci. 39(2):
498-505.
Ruppert, E. E., and P. R. Smith. 1988. The functional organization of
hltration nephridia. Biol. Rev 63: 231-258.
Ruthensteiner, B., and K. Schaefer. 1991. On the protonephndia and
'larval kidneys' of Xassarius (Hinia) reticulattis (Linnaeus) (Caen-
ogastropoda). J. Mull. Stud. 57: 323-329.
Sardet, C., M. Pisam, and J. Maetz. 1979. The surface epithelium of
teleostean fish gills. Cellular and junctional adaptations of the chloride
cell in relation to salt adaptation. / Cell Biol. 80: 96-1 17.
Schmid, R. W., and C. N. Reilley. 1957. New complexon for titration
of calcium in the presence of magnesium. Anal. Chcm. 29: 264-268.
Skelding, J. M., and P. F. Newell. 1975. On the functions of the pore
cells in the connective tissue of terrestrial pulmonate molluscs. Cell
Tissue Res. 156: 381-390.
Sminia, T.. and H. II. Boer. 1973. Haemocyanin production in pore
cells of the snail Lymnaea stagnalis. Z. Zellforsch. 145: 443-445.
Smith, P. R., and E. E. Kuppert. 1988. Nephridia. In Ultrastructure of
the Polvchaeta W. Westheide and C. O. Hermans, eds. Microfauna
Manna 4: 231-262.
Stockmann-Bosbach, R. 1988. Early stages of the encapsulated devel-
opment of Nucel/a lapillus (Linnaeus) (Gastropoda. Muricidae). J.
Moll Stud. 54: 181-1%.
Stockmann-Bosbach, R., and J. Althoff. 1989. A correlated morpho-
logical and biochemical study of capsular fluid of Nucella lapillus
(Gastropoda: Prosobranchia: Muricidae). Mar. Biol. 102: 283-289.
Strangvtays-Dixon, J., and D. S. Smith. 1970. The fine structure of
gill 'podocytes' in Panulints argnx (Crustacea). Tissue Cell 2:
611-624.
Taylor, H. H., and P. Greenway. 1979. The structure of the gills and
lungs of the arid-zone crab, Holthuisana (Austrothelphusa) transversa
(Brachyura: Sundathelphusidae) including observations on arterial
vessels within the gills. J. Zool.. Land 189: 359-384.
Thorpe, E. 1930. A Dictionary of Applied Chemistry. Longmans, Green
and Co.. London.
Tomlinson, S. G. 1987. Intermediate stages in the embryonic devel-
opment of the gastropod Ilyanassa obsoleta: a scanning electron mi-
croscope study. Int. J Invert. Reprod Dev 12: 253-280.
Wessing, A., and A. Polenz. 1974. Structure, development and function
of the protonephridia in trochophores of Pomatoceros triqueter(An-
nelida, Polvchaeta, Sedentaria). Cell Tissue Res. 156: 21-33.
Wilson, R. A., and L. A. Webster. 1974. Protonephndia. Biol. Rev.
49: 127-160.
Wood, R. L., and J. H. Luft. 1965. The influence of buffer systems on
fixation with osmium tetroxide. J. Ultrastruct. Res. 12: 22-45.
Wright, K. A. 1964. The fine structure of the nephrocyte of the gills
of two marine decapods. / Ultrastruct. Res 10: 1-13.
Reference: Biol. Bull 182: 324-332. (June, 1992)
Temperature Stress Causes Host Cell Detachment
in Symbiotic Cnidarians: Implications
for Coral Bleaching
RUTH D. GATES, GAREN BAGHDASARIAN, AND LEONARD MUSCATINE
Department of Biology, University of California, Los Angeles, California 90024
Abstract. During the past decade, acute and chronic
bleaching of tropical reef corals has occurred with in-
creasing frequency and scale. Bleaching, i.e., the loss of
pigment and the decrease in population density of sym-
biotic dinoflagellates (zooxanthellae), is often correlated
with an increase or decrease in sea surface temperature.
Because little is known of the cellular events concomitant
with thermal bleaching, we have investigated the mech-
anism of release of zooxanthellae by the tropical sea ane-
mone Aiptasia pulchella and the reef coral Pocillopora
damicornis in response to cold and heat stress. Both spe-
cies released intact host endoderm cells containing zoox-
anthellae. The majority of the released host cells were
viable, but they soon disintegrated in the seawater leaving
behind isolated zooxanthellae. The detachment and re-
lease of intact host cells suggests that thermal stress causes
host cell adhesion dysfunction in these cnidarians.
Knowledge of the cellular entity released by the host dur-
ing bleaching provides insight into both the underlying
release mechanism and the way in which natural envi-
ronmental stresses evoke a bleaching response.
Introduction
Most tropical corals and sea anemones (Phylum Cni-
daria) contain large populations of symbiotic dinoflagel-
lates (zooxanthellae). The zooxanthellae are located in
vacuoles within the host endoderm cells (Glider ct a/..
1980; Trench. 1987) where they mediate the flux of carbon
and nutrients between the host and the environment
(Muscatine, 1990).
Zooxanthellae-cnidarian symbioses are normally stable;
that is, they have a relatively constant ratio of zooxan-
Received 17 January 1992; accepted 17 March 1992.
thellae to host biomass (Drew, 1972). During the past
decade, however, ecologists have observed that relatively
small changes in the physical parameters of the marine
environment can dramatically influence the stability of
these symbioses (Glynn, 1990). Cnidarian bleaching and
mortality have often been correlated with unusually high
or low sea surface temperatures in tropical oceans world-
wide (Brown and Suharsono. 1990; Coles and Fadlallah,
1990; Glynn, 1990; Williams and Bunkley- Williams,
1990). Bleaching has been attributed to a reduction in the
amount of chlorophyll a (Coles and Jokiel, 1 977; Kleppel
ct a/.. 1989; Porter ct al., 1989; Szmant and Gassman,
1990) and accessory pigments (Kleppel el al, 1989) per
zooxanthella cell, a decline in the population density of
the zooxanthellae (Fisk and Done, 1985; Hoegh-Guldberg
and Smith, 1989), or both (Glynn and D'Croz, 1990;
Lesser el al., 1990). Loss of zooxanthellae perse has been
described extensively at the organismic level (Jaap. 1979;
Gates, 1990; Glynn and D'Croz, 1990; Goreau and Mac-
farlane, 1990; Hayes and Bush. 1990; Jokiel and Coles,
1990; Lesser ct al., 1990; Szmant and Gassman, 1990),
yet few investigators have addressed the underlying cel-
lular mechanism (see O'Brien and Wyttenbach, 1980;
Sandeman. 1988; Lesser ct al.. 1990) or the morphology
of the cellular entity released. Insight into these features
is essential for an understanding of how sea surface tem-
perature anomalies or other environmental stresses de-
stabilize zooxanthellae-cnidarian symbioses.
Zooxanthellae could be released by any of five mech-
anisms (Fig. 1), four of them resulting in the release of
morphologically characteristic cellular entities. The five
mechanisms are: (a) exocytosis of zooxanthellae from the
host cell, resulting in the release of isolated algae (Steen
and Muscatine, 1987); (b) apoptosis (programmed cell
death) and (c) necrosis, both resulting in the release of
324
CELL DETACHMENT IN SYMBIOTIC CNIDARIANS
325
Host Cell Endoderm
Cellular Product
Mechanism
EXOCYTOSIS
APOPTOSIS
NECROSIS
PINCHING OFF
HOST CELL
DETACHMENT
Figure 1. A schematic representation of five potential mechanisms
by which zooxanthellae could be released from the endoderm of cni-
danans. and the cellular entities associated with each mechanism, m,
mesoglea; vm, host vacuolar membrane; hn, host cell nucleus; zx, zoox-
anthella (shaded for clarity of presentation).
zooxanthellae associated with remnants of the host cell
(Searle a ai, 1982); (d) pinching off of the distal portion
of the host cell, resulting in the release of zooxanthellae
surrounded by the vacuolar and pinched off plasma
membrane (Glider, 1983); and (e) detachment of endo-
derm cells from the host and release of these intact cells
containing their complement of zooxanthellae.
Because cnidarians can be readily bleached in the lab-
oratory by brief exposure to low (Steen and Muscatine,
1987; Muscatine el ai, 1991) or high (Hoegh-Guldberg
and Smith, 1989; Glynn and D'Croz, 1990) seawater
temperature, the mechanism of bleaching and the mor-
phology of the cellular entities released can be investigated
experimentally. This approach, together with scanning
electron microscopy of endoderm of the Hawaiian sea
anemone A. pulchella after experimental cold shock, re-
vealed profiles that were interpreted as evidence of exo-
cytosis of zooxanthellae (Steen and Muscatine, 1987). In-
deed, examination of the cellular entities released 24 h
after cold stress revealed abundant isolated zooxanthellae.
In this paper, we describe the cellular entity released
by A. piilclielld and the Hawaiian coral Pocillopora dam-
icornis immediately after a brief exposure to low or high
temperature. P. damicornis is one of several coral genera
that have undergone extensive bleaching in the tropical
eastern Pacific during the elevated temperature of the El
Nino-southern oscillation event (Glynn, 1990), and during
upwelling and seasonal low temperatures (see Discussion
in Glynn and D'Croz, 1990; see also Walker el a!., 1982).
Both species can be bleached in the laboratory, and
bleaching is due to a reduction in zooxanthellae popu-
lation density (Glynn and D'Croz, 1990; Muscatine el al.,
1991). Observations of the cellular entities released at
hourly intervals during, and shortly after, both cold and
heat stress, showed clearly that temperature stress causes
detachment and release of intact endoderm cells contain-
ing zooxanthellae. Soon after release, the host cells dis-
integrate in the environment, leaving isolated zooxan-
thellae.
Materials and Methods
Animal collection and maintenance
A. pulchella and P. damicornis were collected at 1 meter
depth on Checker Reef adjacent to the Hawaii Institute
of Marine Biology (HIMB), Coconut Island, Oahu, Ha-
waii. Habitat temperatures range annually from 21-22°C
to 26-27°C (Jokiel and Coles, 1977). P. damicornis col-
onies were placed in running seawater and used for ex-
periments at HIMB within three days of collection. A.
pulchella was transported to the University of California
at Los Angeles, maintained in an aquarium at 25 °C on
a 1 2 h light/dark regime, and fed twice a week on Anemia
nauplii. Prior to experiments, the anemones were starved
for 24 h in an incubator (Precision Scientific Model 8)
at 25°C on a 12 h light/dark cycle at 40 ^mole
quanta irr2'S~'.
Temperature treatments
All experiments were carried out in darkness following
the protocol of Muscatine et al. (1991).
Cold stress. Individuals of A. pulchella were incubated
in Petri dishes (35 X 10 mm) containing 4 ml of 0.45 ^m
Millipore filtered seawater (MFSW) chilled to 12°C. After
2.5 h, the chilled seawater was removed and replaced with
seawater at 25°C. Anemones were maintained at 25°C
in an incubator in darkness for 14 h. The cellular entities
released to the seawater were then collected and processed
as described below.
Small branches of P. damicornis (2-3 cm length) were
removed from each coral colony and placed in beakers
containing 25 ml of MFSW chilled to 1 2 °C (for the protein
assay, corals were cold stressed at 14°C). After 4 h, the
326
R. D. GATES ET AL
Figure 2. Left panel: photomicrographs of the host cells released to
the seawater by Aiptasia pulchella in response to cold stress, stained for
viability with fluorescein diacetate (X4000). Right panel: photomicro-
graphs of the host cells released to the seawater by Pocillopnra damicornis
in response to cold stress, stained with the DNA specific fluorochrome
Hoechst 33258 (X4000).
branches were immediately transferred to beakers con-
taining 25 ml of MFSW at ambient temperature (23-
24°C). The beakers were placed in the seawater tables for
ambient temperature control and the coral tissue and sea-
water in the beakers was sampled after 1 2 h. Controls for
both species were treated identically to experimental an-
imals but were maintained at ambient seawater temper-
ature (25°C for A. pulchella and 23-24°C for P. dami-
cornis) for 16-16.5 h.
Heat stress. Individuals of A. pulchella were placed in
Petri dishes (35 X 10 mm) containing 4 ml of MFSW
warmed to 32 °C. Small branches of P. damicornis (2-3
cm length) were placed in beakers containing 25 ml
MFSW pre-heated to 32°C. The animals were maintained
at this temperature for up to 16 h. The water surrounding
experimental specimens was examined microscopically
at hourly intervals and the cellular entity released to the
seawater removed and treated as described below. Control
animals of both species were maintained at ambient sea-
water temperature (25°C for A. pulchella and 23-24°C
for P. damicornis) over the experimental time period.
Staining and epijluorescence microscopy
The cellular entities released during and after temper-
ature stress were collected with a fine bore mouth suction
pipette and deposited onto coverslips coated with poly-
L-lysine (0.1% in distilled water). The entities were stained
for viability with the fluorogenic dye fluorescein diacetate
(Sigma Chemical Co., stock solution 1 5 mg/ml in acetone;
working solution 0.04 ml in 9.96 ml 0.1 M sodium phos-
phate, 3% sodium chloride, 0.004% calcium chloride, pH
7.4). The coverslips were rinsed twice in phosphate buffer,
mounted and viewed under epifluorescence with an
Olympus BH-2 microscope. Non-specific esterases in vi-
able cells hydrolyze non-polar fluorescein diacetate to po-
lar molecular fluorescein (Schupp and Erlandsen, 1987).
Additional coverslips were treated for 30 min with the
DNA specific fluorochrome Hoechst 33258 (Reynolds el
ai. 1986; Sigma Chemical Co., stock solution 5 mg/ml
in distilled water; working solution 0.04 ml stock in 9.96
ml 0. 1 M sodium phosphate, 3%' sodium chloride, 0.004%
calcium chloride). The coverslips were dipped in phos-
phate buffer, mounted, and viewed with epifluorescence
microscopy.
Maceration and electron microscopy
The cellular entities released after cold stress were com-
pared to isolated endoderm cells obtained by maceration
of control anemones and corals. A. pulchella tissue was
macerated using 0.05% collagenase (Type 1, Sigma
Chemical Co.) and P. damicornis tissue was dissociated
using calcium-free artificial seawater (Gates and Musca-
tine, 1992). Endoderm cells released by maceration and
the cellular entities released to the seawater as a result of
temperature stress were collected with a mouth pipette
and transferred onto poly-L-lysine coated coverslips. The
coverslips were immersed in 3% glutaraldehyde in 0.1 M
sodium cacodylate buffer (pH 7.4) for 1 h, rinsed twice
in 0.1 M sodium cacodylate buffer and post-fixed in 1%
osmium tetroxide in 0.1 M sodium cacodylate for 30
min. After dehydration in 30, 50, 70, 90. 95, and 100%
(X3) ethanol. the coverslips and attached cells were im-
mersed in hexamethyldisilazane (Applied Sciences, Inc.)
for 5 min (Nation, 1983), dried in air, and then mounted
on aluminum stubs. The stubs were coated with gold and
viewed on a Cambridge 360 scanning electron microscope,
with an accelerating voltage of 10 kV.
For transmission electron microscopy, the endoderm
cells released by maceration of P. damicornis tissue and
the cellular entities released as a result of temperature
stress were collected and centrifuged (Eppendorf model
5414, full speed for 30 s) in microfuge tubes (Gilson, 1.5
ml). The pellets were fixed as described for scanning elec-
tron microscopy. After partial dehydration by sequential
30 min treatments in 30, 50, and 70% ethanol, the 70%
ethanol was drained from the tube and immediately re-
placed with 2% agar. After the agar solidified, the tube
was cut away from the agar plug containing either cold-
CELL DETACHMENT IN SYMBIOTIC CNIDARIANS
327
Figure 3. Scanning electron micrographs of individual host cells released by Aiplusia pulchella (A) and
Pocillopora damicornis (B) in response to cold stress, and those obtained from .-1 pulchella (C) and P.
damicornis (D) by tissue maceration. Bar = 1 urn.
stressed or macerated cells and dehydration completed
through 90, 95, and 100% (X3) ethanol. The preparations
were embedded via propylene oxide into epoxy resin
(Spurr). Thin sections were cut using a Sorvall 6000 ul-
tramicrotome, stained with lead acetate, and viewed on
a JEOL transmission electron microscope.
Protein determination
To investigate the loss of animal protein to the seawater
as a result of temperature stress, the seawater was removed
from the Petri dishes of cold stressed and control A. pul-
chella and homogenized in a teflon-glass tissue grinder.
Sodium dodecyl sulphate (SDS, 1% in seawater) was added
to each homogenate to a final concentration of 0.05%
(modified from McAuley, 1986). For P. damicornis a 4
ml sub-sample was removed from 25 ml seawater samples,
homogenized, and treated with SDS as described above.
Each sample was incubated at room temperature for 45
min to solubilize protein in the seawater and host cell
membranes associated with released algae. The algae were
pelleted by centrifugation (Damon/IEC model HN-S for
4 min at 3000 rpm) and the supernatant put aside for
protein analysis as described below. Each algal pellet was
resuspended in a known volume of MFSW and the total
number of algae assessed using a hemacytometer.
Two 1 ml samples were removed from each supernatant
and the amount of protein assessed spectrophotometri-
cally using the method of Hartree (1972). To ensure that
protein in the seawater samples was animal in origin and
328
R. D. GATES ET AL
,
*
V
Figure 4. Transmission electron micrographs of host cells released by Pocillopura damicornis in response
to cold stress (A), and tissue maceration (B). HN, host cell nucleus; ZX, zooxanthella; VM. vacuolar membrane;
PM. host cell plasma membrane; and M. mitochondria; Bar = 1 ^m.
not secreted by the algae during the 16.5-h experimental
period, control algae were isolated from anemones using
homogenization and centrifugation. The resulting algal
pellets were washed twice in MFSW and treated for 45
min with 0.05% SDS to solublize any animal protein as-
sociated with the algal cells. After two more washes and
re-suspension in MFSW, the number of algae present was
assessed using a hemacytometer. Algal suspensions were
cold stressed (with controls) as described for whole ani-
mals. SDS was added to a final concentration of 0.05%
and the samples were left at room temperature for 45
min. The algae were removed by centrifugation and
counted again to determine if cells had lysed during the
incubation. The remaining supernatant was assayed for
protein as before.
Results
The entities released during and after temperature stress
appeared to be intact host cells. Those released after cold
stress settled at the bottom of the container. In contrast,
those released after heat stress accumulated at the surface
of the water. Unlike the former, the latter were extremely
difficult to collect and handle. They were too fragile to
manipulate for electron microscopy, but we were able to
view them by epifluorescence microscopy. After staining
with fluorescein diacetate and Hoechst 33258, these cells
were identical in profile to those released after cold stress.
Host cells released in both cases appeared to be viable,
with fluorescence restricted to the narrow compartment
of the host cell cytoplasm that surrounded from one to
five zooxanthellae (Fig. 2, left panel). Fluorescein diacetate
was either not taken up by the zooxanthellae, or it was
taken up but masked by the intense red autofluorescence
of the zooxanthellae chlorophylls and the yellow autoflu-
orescence of the zooxanthellae accumulation bodies.
Staining with the bisbenzamide dye Hoechst 33258 re-
vealed a single nucleus within each of these cells (Fig. 2,
right panel).
When viewed with scanning electron microscopy, the
cells released as a result of low temperature stress exhibited
a morphology that was similar to endoderm cells released
from both P. damicornis and A. pulchella by maceration
(Fig. 3). In both cases, the host cell nucleus was visible
under the plasma membrane. This observation suggested
that entities released by thermal stress were intact cells
and not "pinched off" products. Transmission electron
microscopy confirmed the similarity, and clearly revealed
the host cell plasma membrane, the vacuolar membrane
surrounding the zooxanthellae, the host cell nucleus, and
CELL DETACHMENT IN SYMBIOTIC CNIDARIANS
329
Figure 5. Transmission electron micrographs showing degradation of the host cells released by Pocillopora
damicomis in response to cold stress. After dissociation from the epithelium, the host cell plasma membrane
ruptures (A) and the cytoplasmic constituents are free to disperse in the seawater (B). ZX, zooxanthella;
RPM. ruptured host cell plasma membrane; CC. host cell cytoplasmic constituents; IZX, isolated zooxanthella.
Bar = I nm.
mitochondria (Fig. 4). Once released as a result of tem-
perature stress, the host cells degraded rapidly. The host
cell plasma membrane ruptured, the cytoplasmic com-
ponents dispersed, and the vacuolar membrane disap-
peared completely, leaving isolated algae in the seawater
(Fig. 5).
The release of intact host cells by A. pu/c/ie/la and P.
damicomis after thermal stress was further indicated by
a significant positive correlation between the number of
algae released and the total soluble protein detected in
the surrounding medium after the host cells disintegrated
(Fig. 6). Zooxanthellae and soluble protein released by
unstressed control animals was modest (A. pulc/iella) or
negligible (P. damicomis). Protein released by isolated
zooxanthellae was below the limits of detection, and cell
counts confirmed that isolated zooxanthellae had not lysed
during the incubation (data not shown).
Discussion
The results of this investigation show that transient low
and high temperature stress in darkness causes a reduction
in the population density of zooxanthellae in A. pu/chcl/a
and P. damicomis. Quantitative aspects of this reduction
are described elsewhere (Steen and Muscatine, 1987;
Muscatine et ai, 1991). This reduction is caused largely
by detachment of host cells containing zooxanthellae. The
profiles observed by Steen and Muscatine (1987), and in-
terpreted as exocytosis of zooxanthellae, may have been
incidentally evoked by low temperature, or by other stim-
uli, but neither exocytosis. apoptosis. necrosis, nor pinch-
ing off appear to be primary mechanisms of thermal
bleaching by the cnidarians observed in this investigation.
Loss of host cells may explain why investigators observe
loss of protein by bleached corals in excess of that ac-
counted for by loss of zooxanthellae alone (Porter et al.,
1989; Glynn and D'Croz, 1990; Szmant and Gassman,
1 990). Despite loss of cells, the hosts survive the treatment.
Release of zooxanthellae appears to be a two-phase
process. Time-lapse video of A. pulchella during and after
low temperature shock reveals that host cells containing
zooxanthellae first dissociate from the endoderm and ac-
cumulate in the coelenteron where they form pellets or
remain as loose cells. Then, during the rewarming period,
the pellets and cells are periodically propelled by cilia and
muscles through the actinopharynx to the external me-
dium (Hoegh-Guldberg, 1989; Muscatine et al., 1991 ). A
protocol using gradual change in temperature also re-
vealed host cell detachment. However, this protocol was
dismissed in favor of the precipitous change in tempera-
ture because the former required a more lengthy and
complex sampling regime.
We speculate that the dissociation of host cells from
the endoderm is caused by host cell adhesion dysfunction.
330
R. D. GATES ET AL.
1 2 3
No. zooxanthellae released (x l(f )
450
400
350
I 3°
3>
<D 250 -
Q.
V 15
a
I 10°
50 -i
Pocillopora damicornis
02468
No. of zooxanthellae released (x 106)
Figure 6. Appearance of soluble protein in the incubation medium
concomitant with release of zooxanthellae by Aipiasia pulchella and
Pocillopora damicornis. Control (open squares), cold stressed (closed
squares). Line fit with linear regression (Zar, 1984), for A. pulchella,
r = 0.81 (y = 6.6012 + (2.71 • 10~5)x). For P. damicornis. r = 0.96
(y = 10.2124 + (6.28-lQ-5)x).
The effect of high and low temperature stress on cell
adhesion and cytoskeletal organization has been investi-
gated extensively in other systems. Cell adhesion dys-
function may result from temperature-induced membrane
thermotropism (Melchior and Steim, 1976;Quinn, 1989)
and passive influx of ions (Grisham and Barnett. 1973;
Larsen et ai. 1988), especially calcium which, in turn,
may cause the collapse of actin and the intermediate fil-
aments vimentin and cytokeratin (Van Bergen en He-
negouwen, 1985; Coakley, 1987; Wachsberger and Coss,
1989; Cress et al., 1990; Walter et ai, 1990). Cytoskeletal
elements are co-located with the cytoplasmic domain of
cell adhesion molecules (Hirano ct al.. 1987). As elements
of the cytoskeleton and cell adhesion proteins function
as a whole to maintain the integrity of epithelia, disruption
of the former may cause dysfunction of the latter (Tak-
eichi, 1988). Alternatively, temperature stress may cause
denaturation of proteins involved in cell adhesion (Watson
and Morris, 1987; Suzuki and Choi, 1990).
Although we have described the cellular entities released
after thermal stress in darkness, and a probable underlying
mechanism, low salinity (Goreau, 1964; Egana and Di-
salvo, 1982) and sedimentation (Acevedo and Goenaga,
1986) also evoke bleaching. Moreover, at high tempera-
ture, the bleaching response in some cnidarians is thought
to be exacerbated by high irradiance (Coles and Jokiel,
1978), ultraviolet radiation (Harriot, 1985; Jokiel and
York, 1982; Lesser et ai, 1990), and active oxygen (Lesser
and Shick, 1990). These other types of stress cause de-
creased zooxanthellae population density, but the mech-
anism of bleaching in each instance is still unknown. It
may be fundamentally different from that observed in
thermal bleaching. For example, we speculate that low
salinity may cause the cnidarians to lose zooxanthellae
by the mechanical disruption caused by hypoosmotic
shock (i.e., necrosis). We suggest that bleaching be defined
more rigorously in terms of both the environmental stress,
and the morphology of the cellular entity released. Studies
are now under way to determine if host cell detachment
after thermal stress is a general phenomenon or specific
to selected cnidarian genera.
Acknowledgments
We thank Alicia Thompson and Birgitta Sjostrand for
assisting with electron microscopy, Gordon Grau for pro-
viding an epifluorescence microscope at HIMB, and the
Office of Naval Research (Grant #NOOO14-89-J-3246 to
L.M.) and the National Science Foundation (Grant
#OCE-8723090 to L.M.) for research support.
Literature Cited
Acevedo, R., and C. Goenaga. 1986. Note on coral bleaching after a
chronic flooding in southwestern Puerto Rico. Caribb. J Sci. 22:
225.
Brown, B. E., and Suharsono. 1990. Damage and recovery of coral
reefs affected by El Nino related seawater warming in the Thousand
Islands. Indonesia. Coral Reefs 8: 63-170.
Coakley, W. T. 1987. Hyperthermia effects on the cytoskeleton and
on cell morphology. Soc. Exp. Biol. 41: 187-21 1.
Coles, S. J., and Y. H. Fadlallah. 1990. Reef coral survival and mortality
at low temperatures in the Arabian Gulf: new species-specific lower
temperature limits. Coral Reefs 9: 231-237.
Coles, S. L., and P. L. Jokiel. 1977. Effects of temperature on pho-
tosynthesis and respiration in hermatypic corals. Mar. Biol 43: 209-
216.
Coles, S. L., and P. L. Jokiel. 1978. Synergistic effects of temperature.
salinity and light on the hermatypic coral Monnpora vermcosa. Mar
Biol. 49: 187-195.
Cress, A. E., J. A. Majda, J. R. Glass, D. E. Stringer, and E. W. Gerner.
1990. Alteration of cellular adhesion by heat shock. Exp. Cell Res
190: 40-46.
CELL DETACHMENT IN SYMBIOTIC CNIDARIANS
331
Drew, E. A. 1972. The biology and physiology of alga-invertebrate
symbioses. II. The density of symbiotic algal cells in a number of
hermatypic corals and alcyonanans from various depths. / Exp. Mar.
Biol Ecol 9:71-75.
Egana. A. C., and L. H. DiSalvo. 1982. Mass expulsion of zooxanthellae
by Easter Island corals. Pac Set. 36: 61-63.
Fisk, T. A., and T. J. Done. 1985. Taxonomic and bathymetnc patterns
of bleaching in corals. Myrmidon Reef (Queensland). Proc. Fifth Inl
Coral Reef Cong. 6: 149-154.
Gates, R. D. 1990. Seawater temperature and sublethal coral bleaching
in Jamaica. Coral Reefs 8: 193-198.
Gates, R. D., and L. Muscatine. 1992. Three methods for isolating
viable anthozoan endoderm cells with their intracellular symbiotic
dinoflagellates. Coral Reefs (in press).
Glider, W. V. 1983. The biology of the association of Symbiodiniitm
microadriaticum with Aiptasia pal/ida: an anemone-alga symbiosis.
Ph.D. Thesis, University of Nebraska. 102 pp.
« Glider, W. V., D. Phipps, and R. L. Pardy. 1980. Localization of sym-
biotic dinoflagellate cells within tentacle tissue of Aiptasia pallida
(Coelenterata, Anthozoa). Trans. Am. Microsc. Soc. 99: 426-438.
Glynn, P. 1990. Coral mortality and disturbances to coral reefs in the
tropical eastern Pacific. Pp. 55- 1 26 in Global Ecological Consequences
of the 1982-83 El-Nino Southern Oscillation, P. W. Glynn, ed. El-
sevier, Amsterdam.
Glynn, P., and L. D'Croz. 1990. Experimental evidence for high tem-
perature stress as the cause of El Nino-coincident coral mortality.
Coral Reefs %: 181-192.
Goreau, T. F. 1964. Mass expulsion of zooxanthellae from Jamaican
reef communities after Hurricane Flora. Science 145: 383-386.
Goreau, T. J., and A. H. Macfarlane. 1990. Reduced growth rate of
Monlastrea annulans following the 1987-1988 coral-bleaching event.
Coral Reefs 8: 211-216.
Grisham, C. M., and R. E. Barnett. 1973. The role of lipid-phase tran-
sitions in the regulation of the (sodium + potassium) adenosine tri-
phosphatase. Biochemistry 12: 2635-2637.
Harriot, V. J. 1985. Mortality rates of scleractiman corals before
and during a mass bleaching event. Mar. Ecol. Prog. Ser 21:
81-88.
Hartree, E. F. 1972. Determination of protein: a modification of the
Lowry method that gives a linear photometric response. Anal.
Biochem. 48: 422-427.
Hayes, R. L., and P. G. Bush. 1990. Microscopic observations of
recovery in the reef-building scleractmian coral, Monlastrea annu-
lans, after bleaching on a Cayman reef. Coral Reefs 8: 203-
209.
Hirano, S., A. Nose, K. Hatta, A. Kawakami, and M. Takeichi.
1987. Calcium-dependent cell-cell adhesion molecules (Cadherins):
subclass specificities and possible involvement of actin bundles. J.
Cell Biol 105:2501-2510.
Hoegh-Guldberg, O. 1989. The regulatory biology of plant-animal en-
dosymbiosis. Ph.D. Thesis, University of California at Los Angeles.
309 pp.
Hoegh-Guldberg, O., and G. J. Smith. 1989. The effect of sudden
changes in temperature, light, and salinity on population density and
export of zooxanthellae from the reef corals Seriatopora hystrix
and Stvlophora pistillata. J Exp. Mar Biol Ecol 129: 279-
303.
Jaap, \V. 1979. Observations on zooxanthellae expulsion at Middle
Sambo Reef, Florida Keys. Bull Mar Sci 29: 414-422.
Jokiel, P. L., and S. L. Coles. 1977. Effects of temperature on the
mortality and growth of Hawaiian reef corals. Mar Biol. 43: 201-
208.
Jokiel, P. L., and S. L. Coles. 1990. Response of Hawaiian and other
Indo- Pacific reef corals to elevated temperature. Coral Reefs 8: 1 55-
162.
Jokiel, P. L., and R. H. York. 1982. Solar ultraviolet photobiology of
the reef coral Pocillopora damicornis and symbiotic zooxanthellae.
Bull. Mar. Sci. 32: 301-315.
Kleppel, G. S., R. E. Dodge, and C. J. Reese. 1989. Changes in pig-
mentation associated with the bleaching of stony corals. Limnol
Oceanogr 34: 1331-1335.
Larsen, T., S. Solberg, R. Johansen, and L. Jorgensen. 1988. Effect of
cooling in the intracellular concentrations of Na+, K* and Cl~ in
cultured human endothelial cells. Scand. J Lab. Invest. 48: 565-
571.
Lesser, M. P., and J. M. shirk 1990. Photoadaptation and defenses
against oxygen toxicity in zooxanthellae from natural populations
of symbiotic cnidarians. J. Exp. Mar. Biol. Ecol. 134: 129-
141.
Lesser, M. P., W. R. Stochaj, D. W. Tapley, and J. M. Shick.
1990. Bleaching in coral reef anthozoans: effects of irradiance, ul-
traviolet radiation, and temperature on the activities of protective
enzymes against active oxygen. Coral Reefs 8: 225-232.
McAulcy, P. J. 1986. Isolation of viable uncontaminated Chlorella
from green hydra. Limnol. Oceanogr. 31: 222-224.
Melchior, D., and J. M. Steim. 1976. Thermotropic transitions in bio-
membranes. Ann. Rev. Biophys. Bioeng. 5: 205-238.
* Muscatine, L. 1990. The role of symbiotic algae in carbon and energy
flux in reef corals. Pp. 75-87 in Coral Reefs, Z. Dubinsky. ed. Elsevier,
Amsterdam.
Muscatine, L., D. Grossman, and J. Doino. 1991. Release of symbiotic
algae by tropical sea anemones and corals after cold shock. Mar.
Ecol. Prog. Ser 77: 233-243.
Nation, J. L. 1983. A new method using hexamethyldisilazane for
preparation of soft insect tissue for scanning electron microscopy.
Stain. Techno/. 56: 347-351.
O'Brien, T. L., and C. R. Wyttenbach. 1980. Some effects of temper-
ature on the symbiotic association between zoochlorellae (Chloro-
phyceae) and the sea anemone Anthopleiira xanthogrammica. Trans.
Am Microsc. Soc. 99: 221-225.
Porter, J. W., W. K. Fitt, H. J. Spero, C. S. Rogers, and M. W. White.
1989. Bleaching in reef corals: physiological and stable isotope re-
sponses. Proc. Nail. Acad. Sci. USA 86: 9342-9346.
Quinn, P. J. 1989. Principles of membrane stability and phase behavior
under extreme conditions. / Bioenergetics Biomembranes 21:
3-19.
Reynolds, C. P., A. T. Black, and J. N. Woody. 1986. Sensitive method
for detecting viable cells seeded into bone marrow. Cancer Res. 46:
5878-5881.
Sandeman, I. M. 1988. Coral bleaching at Discovery Bay, Jamaica: a
possible mechanism for temperature-related bleaching. Pp. 46-48 in
Mass Bleaching of Corals in the Caribbean: A Research Strategy.
Research report 88-2, J. Ogden and R. Wicklund, eds. NOAA Un-
dersea Research Program, Rockville, MD.
Schupp, D. G., and S. L. Erlandsen. 1987. A new method to determine
Giardia cyst viability: correlation of fluorescein diacetate and pro-
pidium iodide staining and animal infectivity. Appl. Environ. Micro-
biol. 53: 704-707.
Searle, J., J. F. R. Kerr, and C. J. Bishop. 1982. Necrosis and apoptosis:
distinct modes of cell death with fundamentally different significance.
Pathol. Annu. 17: 229-259.
Steen, R. G., and L. Muscatine. 1987. Low temperature evokes rapid
exocytosis of symbiotic algae by a sea anemone. Biol. Bull 172: 246-
263.
Suzuki, M., and B. H. Choi. 1990. The behavior of the extracellular
matrix and the basal lamina during the repair of cryogenic injury1
in the adult rat cerebral cortex. Ada. Neiiropalhol. 80: 355-
361.
332
R. D. GATES ET AL
Szmant, A., and N. J. Gassman. 1990. The effects of prolonged
"bleaching" on the tissue biomass and reproduction of the reef coral
Montasirea annularis. Cora/ Reefs 8: 2 1 7-224.
Trench, R. K. 1987. Dinoflagellates in non-parasitic symbiosis. Pp. 530-
570 in Biology of Dinoflagellates, F. J. R. Taylor, ed. Blackwell, Ox-
ford.
Takeichi, M. 1988. The cadherins: cell-cell adhesion molecules con-
trolling animal morphogenesis. Development 102: 639-655.
Van Bergen en Henegouwen, P. M. P., W. J. R. M. Jordi, G. Van Dongen,
F. C. S. Ramaekers, H. Amesz, and W. A. M. Linnemans.
1985. Studies on a possible relationship between alterations in the
cytoskeleton and induction of heat shock protein synthesis in mam-
malian cell. Ini ./. Hyperthermia 11: 69-83.
Wachsberger, P. R., and R. A. Coss. 1989. Acrylamide sensitization
of the heat response of the cytoskeleton and cytotoxicity in attaching
and well-spread synchronous Chinese hamster ovary cells. Cell Mo-
tiliiy Cytoskel. 13: 67-82.
Walker, N. D., H. H. Roberts, L. J. Rouse, Jr., and O. K. Huh.
1982. Thermal history of reef-associated environments during a
record cold-air outbreak event. Coral Reefs 1: 83-87.
Walter, M. F., N. S. Petersen, and H. Biessman. 1990. Heat shock
causes the collapse of the intermediate filament cytoskeleton in Dro-
sophila embryos. Dev. Genet. 11: 270-279.
Watson, P. F., and G. J. Morris. 1987. Cold shock injury in animal
cells. Symp. Soc. E.\p. Biol. 41: 31 1-340.
Williams, E. H., Jr., and L. Bunkley-Williams. 1990. The world-wide
coral reef bleaching cycle and related sources of coral mortality. Atoll
flf.v. Bull. 335: 1-71."
Zar, J. H. 1984. Biostatistical Analysis, 2nd ed. Prentice Hall, New
Jersev.
Reference: Biol. Bull 182: 333-340. (June. 1992)
Modulation of Crayfish Hearts
by FMRFamide-related Peptides
A. JOFFRE MERCIER1 AND RUNE T. RUSSENES
Department oj Biological Sciences. Brock University. St. Catharines. Ontario L2S 3 A 1 Canada
Abstract. The present study examines effects of FMRF-
amide-related peptides (FaRPs) on crayfish heart. Lobster
peptides F, (TNRNFLRFamide) and F2 (SDRNFLRF-
amide) increase the rate and amplitude of heart beat in
hearts isolated from Procambarus clarkii. Thresholds for
these effects were between 10~'° and 1CT9 M for F2 and
between 1(T9 and 10~8 M for F,. FMRFamide and
FLRFamide elicited similar cardioexcitatory effects, but
at thresholds of approximately 1CT7 M. Thus, the amino-
terminal extensions "TNRN" and "SDRN" enhance the
excitatory actions of FMRFamide and FLRFamide.
SchistoFLRFamide (PDVDHVFLRFamide) and leu-
comyosuppressin (pQDVDHVFLRFamide) markedly
decrease the rate of cardiac contractions at 1CT9 to 10~8
M and can suppress the cardiac rhythm for one minute
or more at 1CT7 M. The amino-terminal extensions of
these two peptides, therefore, are necessary for inhibition
of heart rate. Both of these peptides cause an initial re-
duction in contraction amplitude, but contractions sub-
sequently increase in the presence of SchistoFLRFamide.
Thus, crayfish hearts are sensitive to several FMRFamide-
related peptides, but the sites and mechanisms of action
remain to be determined.
Introduction
Since the discovery of the neuropeptide FMRFamide
(Phe-Met-Arg-Phe-NH:) in the bivalve mollusk Macro-
callista nimbosa (Price and Greenberg. 1977). FMRFam-
ide-related peptides (FaRPs) have been reported in nu-
merous invertebrate and vertebrate species (e.g.. Boer et
ai, 1980; Dockray et al, 1983; Watson el al., 1984:
Grimmelikhuijzen and Graff, 1985; Lehman and Price.
Received 16 September 1 99 1 ; accepted 26 March 1992.
1 To whom correspondence should be addressed.
1987; Li and Calabrese. 1987; Elphick el al.. 1989: Robb
et al.. 1989; Krajniak and Price, 1990). It is now recog-
nized that FaRPs represent members of a very large family
of neuropeptides that is widely distributed throughout the
animal kingdom (Greenberg and Price, 1983; Price and
Greenberg. 1989).
In crustaceans. FMRFamide-like immunoreactivity
(FLI) has been found throughout most of the nervous
system, but the highest amounts are present in the peri-
cardial organs (POs) (Koberski et al., 1987; Marder et al.,
1987; Krajniak, 1991: Mercier el al.. 1 99 Ib). Because the
POs are located in the pericardia! sinus just outside the
heart (Maynard, 1960). and because they release car-
dioactive hormones (e.g.. Cooke and Sullivan. 1982;
Kravitz el al.. 1980), the heart is likely to be an important
target for crustacean FaRPs. So far, only two FaRPs have
been sequenced and identified in extracts of the lobster
POs, although other FMRFamide-like immunoreactive
material is present (Trimmer et al.. 1987). These two pep-
tides have the sequences TNRNFLRFamide (F,) and
SDRNFLRFamide (F:). Both of these peptides excite iso-
lated hearts of lobsters (Kravitz et al.. 1987) and blue
crabs (Krajniak. 1991 ). However, the effects of these and
other FaRPs on crustacean hearts have not been thor-
oughly investigated.
The primary aim of the present study was to examine
in greater detail the cardio-regulatory effects of lobster
peptides F, and F: . The effects of these two peptides were
studied on isolated crayfish hearts. To examine the rela-
tionship between amino acid sequence and biological ac-
tivity, the effects of F, and F2 were compared with those
of FMRFamide. FLRFamide. and two FaRPs with sig-
nificantly different amino-terminal extensions: leuco-
myosuppressin (LMS). with the sequence pQDVDHV-
FLRFamide (Holman el ai. 1986) and SchistoFLRF-
amide (Sch). with the sequence PDVDHVFLRFamide
333
334
A J. MERCIER AND R. T RUSSENES
(Robb el al.. 1989). The crayfish heart was sensitive to all
of the compounds tested.
Materials and Methods
Crayfish were obtained commercially and were main-
tained in aerated freshwater tanks at 14.5°C on a mixed
vegetable diet.
Synthetic peptides were applied to spontaneously active
crayfish hearts. The dorsal carapace, containing the heart
and pericardium, was dissected from crayfish weighing
approximately 3 g and was pinned to a Sylgard-lined dish
with the ventral side up. The pericardial membrane was
severed and pinned at each side to allow the bathing fluid
access to the heart. The recording chamber, which had a
volume of 0.5 ml, was perfused with crayfish saline (van
Harraveld, 1936). which had the following constituents
(in mM): Na+, 205; Cr, 232; K+, 5.3; Ca++, 13.5; Mg++,
2.5; HEPES, 5.0 (pH 7.4). Saline was added to the cham-
ber at a rate of 3.0 ml min^1 using a peristaltic pump and
was removed at the other end of the chamber by suction.
The entire preparation was superfused continuously in
this manner. The temperature was maintained at 14-
16°C, but did not vary by more than 1°C during any one
experiment. Heart preparations were viable for up to 8 h.
Contractions were recorded by connecting the sternal
artery to a tension transducer using two insect pins that
were hooked at one end and glued to the transducer at
the other end. The artery and heart were stretched until
the maximum contraction amplitude was obtained. Con-
tractions were displayed on a Grass Model 7B Polygraph.
The rate and amplitude of contractions were measured
manually over intervals of either 30 s or 1 min.
Peptides were applied by changing the perfusate to a
solution containing a selected peptide concentration.
Peptides were present in the bathing chamber for 8-10
min; 10 min was chosen arbitrarily as the maximum time
of exposure. In a few cases, it was obvious that the max-
imal effect of the peptide had already occurred, and wash-
out was begun after 8 min. The maximal effect of the
peptides generally occurred well before this time, except
for the increased amplitude caused by SchistoFLRFamide
(see Results).
During washout, the effects of the peptides began to
subside within 5 min and had completely worn off within
20-30 min. In many cases (e.g.. Fig. 7), the effects wore
off more rapidly. Each peptide was tested by starting with
the lowest concentration (10~': M) and subsequently in-
creasing the dosage in 10-fold increments until the entire
response range had been tested. Successive doses were
always given after the previous dose was completely
washed out. Three successive applications of 1 X 10~9 M
F: onto the same preparation yielded virtually identical
effects each time. The results reported here were obtained
from a total of 37 preparations. In all but six experiments,
each peptide was tested on one preparation only. The
number of preparations used for testing each peptide is
indicated in the figure legends.
SchistoFLRFamide and leucomyosuppressin were ob-
tained from Peninsula Laboratories Ltd. (Belmont, Cal-
ifornia). Peptides F[ and F2 were synthesized by Dr. D.
McCormack (Rochester, Minnesota) and were a gift from
Dr. M. Schiebe. FMRFamide, FLRFamide, and all other
chemicals were obtained from Sigma Chemical Co. (St.
Louis, Missouri).
In each case, the error value reported is the standard
error of the mean. Statistical significance of differences
between mean values was determined using a Student's t
Test (Furguson, 1971).
Results
Of the six neuropeptides tested, four elicited responses
that were exclusively cardio-excitatory. FMRFamide,
FLRFamide, F, , and F: increased both the rate and am-
plitude of contractions of isolated Procambarus hearts.
Figure 1 shows representative examples of the effects of
these four peptides. at doses that elicited approximately
equivalent responses. The onset of such responses usually
10 -9 M F2
108 M
10 7 M FLRFa
10-6M FMRFa
Figure 1. Effects of excitatory FaRPs on crayfish heart contractions.
Each panel shows chart recordings of spontaneous contractions. Peptides.
at the concentrations indicated, were present in the bathing solution
during the periods indicated by the thick horizontal bars. The recordings
were obtained from different preparations. (Abbreviations: FLRFa for
FLRFamide, FMRFa for FMRFamide, F, and F2 as in text.)
FARPS MODULATE CRAYFISH HEART
335
occurred within 60-90 s after the peptide entered the
bathing chamber, and the lag-time was shorter at higher
peptide concentrations. The effects on heart rate and con-
traction amplitude were always completely reversed by
20-30 min of washing in normal saline (data not shown).
None of these four peptides elicited any inhibitory effects.
Log dose-response curves, constructed for the four ex-
citatory peptides (Fig. 2), were based on the responses of
five to six preparations in each case. Responses were ex-
pressed as the percentage change in contraction rate and
the percentage change in amplitude by comparing the
maximal effect of each peptide with the average contrac-
tion rate, or contraction amplitude, during the 3 min pe-
riod immediately preceding peptide application. Peptides
F, and F2 caused a more pronounced increase in the am-
plitude of the contractions than on their rate. For
FLRFamide and FMRFamide, however, the percentage
change in amplitude was similar to the change in rate in
each case.
Differences in the relative potencies of the peptides were
more prominent for the effect on contraction amplitude
(Fig. 2A) than for the effect on rate (Fig. 2B). A compar-
ison of the effects on contraction amplitude gave the fol-
lowing results. F2 was the most potent peptide, with a
threshold concentration between 10"'° and 10~9 M. A
comparison of the effects of 10~9 M F2 and 10~5 M
FMRFamide suggests that F2 was up to 10,000 times more
potent than FMRFamide. F, was the next most potent
peptide, with a threshold between 10~9 and 10~8 M, and
was approximately 1000 times more potent than
FMRFamide (based on the effects of 10~8 MF, and 10~5
M FMRFamide). FLRFamide was about 10 times more
potent than FMRFamide (based on the effects of 10~7 M
FLRFamide and 10~6 M FMRFamide), but the threshold
concentrations for both appeared to be between 10 8 and
10~7 M. FMRFamide gave a relatively broad log-dose ver-
sus response curve, which rose steadily over the concen-
tration range of 10~8 to 10~5 M, while the other peptides
appeared to reach saturation within slightly narrower
concentration ranges.
Differences in relative potency were not as marked
when comparing the effects of the excitatory peptides on
contraction rate (Fig. 2B). F2 and F, had approximately
equivalent effects on heart rate, but both compounds were
about 100 times more potent than FLRFamide and
FMRFamide. The log-dose versus response curves for
FLRFamide and FMRFamide were very similar.
SchistoFLRFamide (Sell) and leucomyosuppressin
(LMS) had inhibitory effects on cardiac contractions at
concentrations of 10~9 to 10~7 M (Figs. 3-5). The rate of
spontaneous contractions was reduced consistently by
both peptides. At 10 7 M. contractions were completely
suppressed for a period lasting 1 min or longer, after which
-12 -11 -10 -9 -8 -7 -6 -5
log [peptide cone (M)]
B
350l
300
250"
200-
150-
100-
50-
o-
-50
-13 -12 -11 -10 -9 -8 -7 -6 -5 -4
log [peptide cone (M)]
Figure 2. Log-dose versus response curves for the effects of excitatory
FaRPs on the amplitude of cardiac contractions (A) and on heart rate
(B). The percentage change in rate or in amplitude was determined by
comparing the maximum value obtained in the presence of the peptide
from the average value during the three minutes prior to peptide appli-
cation:
% change = [(peak value - initial valuel/initial value] x 100.
Each point represents the mean value for five preparations in the case
of F3 and for six preparations in all other cases. Error bars depict standard
errors of the means. (Abbreviations are as in Fig. 1.)
contractions resumed at a rate that was lower than before
peptide application (Figs. 3, 4, and 6 A). Dose-response
curves, based on the maximal reduction in rate, were
markedly similar for these two peptides (Fig. 5A). The
threshold for inhibition of heart rate was between 10~'°
and 10~9 M.
Effects of Sch and LMS on the amplitude of contrac-
tions were more complex. For most preparations, such as
the one represented in Figure 4, the contractions that per-
sisted in the presence of either Sch or LMS were initially
reduced in amplitude. This type of effect was observed in
4 of 6 preparations exposed to Sch and in 6 of 8 prepa-
rations exposed to LMS. Figure 3 is an example of re-
cordings from a preparation in which no substantial re-
duction in contraction amplitude occurred during expo-
sure to LMS. When observed, reductions in contraction
size were usually transient. Approximately 2-4 min after
peptide exposure began. 10 7 M Sch caused a substantial
increase in contraction size above the level observed prior
336
A. J. MERCIER AND R. T. RUSSENES
10-9 M LMS
10-8 M LMS
10-7 M LMS
JJJjJJjJJJJJJJJM
1 1 rtiN
20 s
Figure 3. Effects of LMS on heart contractions. Each panel shows
chart recordings of spontaneous cardiac contractions. LMS was present
in the bathing solution during the periods indicated by the thick horizontal
bars at the concentrations indicated. The recordings were all from the
same preparation.
to peptide exposure (Fig. 4). Thus, the effect of Sch on
contraction amplitude appeared to be biphasic.
Log-dose versus response curves for the initial effect of
LMS and Sch on contraction amplitude were obtained
by comparing responses 1 min after the peptide entered
the bath with the average amplitude over a 2-min period
prior to peptide exposure (Fig. 5B). (The change in con-
traction size was expressed as a percentage of the ampli-
tude during the period prior to peptide application.) The
dramatic reduction in contraction size at 1(T7 M was due
mainly to the fact that this dose completely suppressed
contractions in most preparations. Log-dose versus re-
sponse curves for the increase in amplitude that occurred
later were obtained by comparing the average contraction
amplitude during the 2-min period before peptide appli-
cation with the highest average amplitude over a 1-min
period during exposure to the peptide or during the first
5 min of the wash-out period (Fig. 5C). On average, con-
traction amplitudes doubled in 10~8 M Sch and tripled
in 1(T7 M Sch. In contrast, 10 7 M LMS caused only a
25% increase in contraction size.
The increase in contraction amplitude caused by Sch
developed more slowly than did the reduction in con-
traction rate. Figure 6 A shows the time course of the effects
of 1(T8 M Sch for an individual preparation. In this case,
heart rate began to decline within 2 min of peptide ex-
posure and reached its lowest value 6 min later. Contrac-
tion amplitude, however, did not begin to rise until 6 min
after the peptide entered the bath and required an addi-
tional 8 min to reach its maximal level. In experiments
with six preparations exposed to 10~7 M Sch, the mean
time for the maximal inhibition of heart rate (1.3 ± 0.36
min) was shorter than the mean time for the maximal
increase in amplitude ( 10.9 ± 2.0 min), and the difference
in means was statistically significant (t = 5.56; P < 0.0 1 ).
In contrast, the increased amplitude caused by F2 oc-
curred rapidly and usually coincided with the maximal
increase in rate, as in the example shown in Figure 6B.
Mean times for the maximal rate (2.7 ± 0.68 min) and
amplitude (3. 1 ± 0.99) for five preparations were not sig-
nificantly different (t = 0.93; P> 0.4). As in the example
shown in Figure 6B, the heart rate often declined rapidly
from the peak value, even though F2 was still present in
the bathing solution. Contraction amplitude, however,
remained elevated until the peptide was removed.
As a first step toward examining potential interactions
between FaRPs, the effect of a mixture of Fi and Sch was
studied. We were particularly interested in determining
10-9 M Sch
10-8 M Sch
10-7 M Sch
|2mN
20 s
Figure 4. Effects of Sch on heart contractions. Each panel shows
chart recordings of spontaneous cardiac contractions. Sch was present
in the bathing solution during the periods indicated by the thick horizontal
bars at the concentrations indicated. The recordings were all from the
same preparation.
FARPS MODULATE CRAYFISH HEART
337
A 3-
-20
-40"
-60-
-80
-100
-13 -12 -11 -10
log [peptide cone (M))
B
20-
0-
-20-
-40-
-60-
-80-
ci.
<
-100
-13 -12 -11 -10 -9 -8 -7
log [peptide cone (M)]
300-1
200-
a.
< o-
S -100
-12 -11 -10 -9 -8 -7
log [peplide cone (M)]
Figure 5. Log-dose versus response curves for effects of LMS and
Sch on heart rate (A), on the amplitude of contractions after I min of
peptide exposure (B) and on the maximum amplitude of contractions
during peptide exposure or in the first 5 min of the wash-out period (C).
The change in rate or in amplitude was determined by comparing the
value obtained in response to the peptide with the average value prior
to peptide application and was expressed as a percentage of the initial
value as in Figure 2. Effects on heart rate (A) were determined using the
minimum heart rate during peptide exposure. Each point represents the
mean value obtained from eight preparations exposed to LMS and from
six preparations exposed to Sch. Error bars depict standard errors of the
whether the chronotropic action of one peptide might
predominate or, alternatively, whether the two substances
might produce an "additive" response. Figure 7 illustrates
the responses of an individual preparation in which 10s
M F: caused a rapid 60%. increase in heart rate and 10 8
M Sch decreased heart rate by about 50%. A mixture of
the two peptides produced a heart rate that increased only
slightly and was comparatively stable. Subsequent appli-
cation of F: showed that the heart was capable of re-
sponding to this peptide as before. Thus, these two pep-
tides exert chronotropic actions that are mutually antag-
onistic and tend to cancel each other out when combined.
Sch did not antagonize the effect of F: on contraction size
and may even have potentiated it (Fig. 7). This experiment
was performed six times with qualitatively similar results
I
« 30-
£ 20-
•
CC
0
I
1 50
•1 25=-
Sch
•075 "°
050 a.
E
•025<
10 20 30
Time (min)
000
40
B
0> 40 -
ra
CC
c 30-
-200
-1 75
-1 50
-1 25
1.00
0.75
-0,50
025
10 15 20 25 30
Time (min)
35
Figure 6. Time course for the effects of Sch (A) and F: (B) on the
rate and amplitude of cardiac contractions. The peptides were present
in the bathing solution at concentrations of 1CT8 M during the periods
indicated by the horizontal bars. The data for (A) and (B) were obtained
from the same experimental preparation.
in each case. Sch always antagonized the chronotropic
effect of F2 but not its inotropic effect. Thus, while F: is
chronotropically and inotropically excitatory, and Sch is
chronotropically inhibitory, a mixture of these FaRPs can
produce a response that is predominantly inotropic.
30
20n
Sch
Sch
50 100
Time (min)
Figure 7. The effect of a mixture of F2 and Sch on the rate (upper
panel) and amplitude (lower panel) of cardiac contractions. The data
were obtained from a single preparation. Peptides were present in the
bathing solution at I O"8 M at the times indicated by the horizontal bars.
338
A. J. MERCIER AND R. T. RUSSENES
Discussion
FaRPs have been reported to modulate the activity of
hearts from several types of invertebrates, including mol-
lusks (e.g.. Painter and Greenberg, 1982; Price et al,
1990), insects (Cuthbert and Evatis, 1989; Robb et al..
1989), leeches (Li and Calabrese, 1987), and crustaceans
(Kravitz et al., 1987; Krajniak, 199 1 ). Some FaRPs elicit
responses that are mainly excitatory, others elicit responses
that are primarily inhibitory, and others are reported to
elicit mixed or biphasic responses (Cuthbert and Evans,
1989). The type of response depends on the chemical
structure of the peptide and on the invertebrate species
(Painter and Greenberg, 1982).
The present study did not investigate the sites or mech-
anisms of action of the various FaRPs tested. Changes in
rate are most likely to result from effects on the cardiac
ganglion, which sets the overall rhythm for cardiac con-
tractions (Maynard, 1960). Changes in the amplitude of
contractions could be due to changes in the number or
frequency of nerve impulses per burst produced in the
cardiac motor neurons (Maynard, 1960), or they could
result from direct effects on the cardiac muscle cells, as
in Limulus (Watson el al., 1985).
It should be noted, however, that the POs, which con-
tain several cardioexcitatory agents (e.g., Cooke and Sul-
livan, 1983), were still present in the preparations used
for study. We cannot exclude the possibility that the ap-
plied peptides act by inducing the release of other agents
from the POs. Such hormone-induced release of hormones
from the same neurosecretory organ has not been reported
for any of the pericardia! substances and seems unlikely.
In addition, the inhibitory effects of Sch and of LMS are
qualitatively similar to those reported in insect hearts,
which lack POs (Cuthbert and Evans, 1989; Robb et al..
1989).
F| and F: exert purely excitatory effects on hearts of
Procambarns, increasing both the rate and amplitude of
spontaneous contractions. F, and F2, respectively, are up
to 1,000 and 10,000 times more potent than FMRFamide
when comparing inotropic responses. Similar results have
been obtained with isolated hearts of the blue crab, Cal-
linecles sapidus (Krajniak, 1991) and of the lobster,
Homarus americanus ( Kravitz et al., 1987). F, is also 1000
times more potent than FMRFamide in exciting the semi-
isolated heart of the locust. Schistocerca gregaria (Cuth-
bert and Evans, 1989). Thus, arthropod hearts appear to
be more sensitive to N-terminally extended analogues of
FLRFamide than to FMRFamide. FMRFamide and
FLRFamide are both excitatory, but the addition of the
amino acid sequence thr-asn-arg-asn- or of ser-asp-arg-
asn- to the amino terminal of the tetrapeptide-phe-leu-
arg-phe-NH2 effectively increases the potency of the pep-
tide. Similar structure-activity relationships have been
reported for the locust extensor tibiae nerve-muscle prep-
aration (Cuthbert and Evans, 1989) and for neuromus-
cular synapses on crayfish abdominal extensor muscles
(Mercierrffl/.. 1990).
The sensitivity of hearts from Homarus (Kravitz et al.,
1987), Callinectes (Krajniak, 1991), and Procambarns to
F, suggests that F, or closely related peptides are common
cardioexcitatory hormones in decapods. F, and F2 were
originally isolated from lobster POs (Trimmer et al.. 1987)
and have not been positively identified in any other tissues
to date. The POs of Procambarns clarkii. however, appear
to contain cardioexcitatory FaRPs that are very similar
to F, and F2 (Mercier et al.. 1991a, b).
The crayfish heart was also sensitive to Sch and LMS,
which have markedly different N-terminal extensions than
the other FaRPs studied. The primary effect of Sch and
of LMS appears to be a reduction in heart rate, which
occurs at a threshold concentration between 10" I0 and
10~9 M for both peptides. Similar thresholds have been
reported for inhibition of locust hearts (Robb et al.. 1989)
and oviducts (Lange et al., 199 1 ) by Sch, and for inhibition
of cockroach hindguts (Holman et a/.. 1986) and locust
hearts (Cuthbert and Evans, 1989) by LMS. Because
FLRFamide excites the crayfish heart, the inhibitory ef-
fects observed in the present study must be due to the
presence of the N-terminal extensions PDVDHV- and
pQDVDHV- on Sch and LMS, respectively.
In contrast, Krajniak (1991) has reported that LMS
causes cardioexcitation in Callinectes. The effect was ob-
served in a single preparation and has a threshold 10,000
times higher than that of FI . Such an observation might
reflect species-dependent differences in the target tissues,
but this possibility requires further study.
Reduction in the amplitude of contractions was not
observed as consistently with crayfish hearts as with insect
preparations (Cuthbert and Evans, 1989; Lange el al.,
1991). The large increase in contraction size at higher Sch
concentrations was distinct from the inhibitory effect in
that it developed much more slowly. In addition, the large
increase in amplitude was only produced by Sch, whereas
both Sch and LMS could inhibit heart rate. This indicates
that Sch activates a receptor that is not activated by LMS.
This might seem surprising, in view of the high degree of
homology between the two peptides, which differ by only
the N-terminal amino acid. It is possible, however, that
at the low concentrations, both peptides activate the same
receptor or receptor class (one responsible for reducing
heart rate), and that a different class of receptor is activated
at higher concentrations of Sch (causing the increase in
amplitude). This hypothesis is consistent with the high
similarity between the log-dose versus response curves for
the effects of Sch and LMS on contraction rate (Fig. 5A).
I \RPS MODULATE CRAYFISH HEART
339
The reason for the delayed inotropic effect of Sch is
not clear. One possibility may be that the peptide is hy-
drolyzed beginning at the N-terminal end. This would
gradually produce a peptide more closely resembling
FLRFamide, which, at 10 7 M, caused a substantial in-
crease in contraction size (Figs. 1, 2). The combination
of Sch with an excitatory FaRP (F2) was capable of in-
creasing the amplitude of contractions without increasing
heart rate (Fig. 7). Thus, if hydrolysis were to cause an
accumulation of an excitatory FaRP, the combination of
Sch and its breakdown product would be expected to pro-
duce a response similar to that observed after several min-
utes of exposure to Sch (Figs. 4, 6A). Such an explanation,
however, is speculative, and others are possible.
Responses to mixtures of excitatory and inhibitory
agents may be of some physiological significance. Increases
in cardiac output (the fluid volume ejected by the heart
per minute) during physical activity (e.g., McMahon et
ul.. 1^79) and hypoxia (e.g.. Burnett, 1979; McMahon
and Wilkens. 1975) generally involve changes in both
heart rate and stroke volume (the amount of fluid ejected
per beat). In some cases, however, stroke volume increases
independently of heart rate (Taylor, 1976; McMahon and
Wilkens, 1977; Taylor and Butler, 1978), and the mech-
anisms that underly such an effect are not known
(McMahon and Burnett, 1990). It is tempting to speculate
that a mixture of peptides like F2 and Sch, which is capable
of increasing the amplitude of contractions independently
of rate, might be involved. Tension recordings, however,
do not provide a direct measure of stroke volume or of
cardiac output. In addition, crayfish POs do not contain
Sch or LMS (Mercier et al. 1991b), and none of the neu-
rohormones isolated from POs to date decreases heart
rate. More research aimed at identifying neuropeptides
and examining their effects may provide a better under-
standing of how stroke volume and cardiac output are
regulated.
Acknowledgments
This work was supported by a grant to AJM from the
Natural Sciences and Engineering Research Council of
Canada. We thank Pat Quigley for assistance in analyzing
data for some of the experiments and Dr. Ian Orchard
for commenting on a draft of the manuscript. Dr. R.
Baines also provided some helpful suggestions.
Literature Cited
Boer, H. H., L. P. C. Schot, J. A. Veenstra, and D. Reichclt.
1980. Immunocytochemical identification of neural elements in
the central nervous systems of a snail, some insects, a fish and a
mammal with an antiserum to the molluscan cardio-excitatory tet-
rapeptide FMRF-amide. Cell Tissue Res. 213: 21-27.
Burnett, L. 1979. The respiratory function of hemocyanin in the spider
crab Lihinia emarginala and the ghost crab Ocypode auadrala in
normoxia and hypoxia. J. Exp. Zool. 210: 289-300.
Cooke, I. M., and R. K. Sullivan. 1982. Hormones and neurosecretion.
Pp. 205-290 in The Biology of Crustacea. D. E. Bliss, H. L. Atwood,
and D. C. Sandeman. eds. Academic Press, New York.
Cuthbert, B. A., and P. D. Evans. 1989. A comparison of the effects
of FMRFamide-like peptides on locust heart and skeletal muscle. J.
Exp. Biol 144: 395-415.
Dockray, G. J., J. R. Reeve, Jr., J. Schively, R. J. Gayton, and C. S.
Barnard. 1983. A novel active pentapeptide from chicken brain
identified by antibodies to FMRFamide. Nature 305: 328-330.
Klphick, M. R., R. H. Emson, and M. C. Thorndyke. 1989. FMRF-
amide-like immunoreactivity in the nervous system of the starfish
Asterias nibens. Biol. Bull. Ill: 141-145.
Furguson, G. A. 1971. Statistical Analysis in Psychology and Education,
McGraw Hill, New York. Pp. 146-159.
Greenberg, M. J., and D. A. Price. 1983. Invertebrate neuropeptides:
native and naturalized. .-I mi. Rev. PhysioL 45: 271-288.
Grimmelikhuijzen, C. J. P., and D. Graff. 1986. Isolation of <Glu-
Gly-Arg-Phe-NH: (AnthoRFamide), a neuropeptide from sea ane-
mones. Proc. Null. ACM/. Sa. USA 83: 9817-9821.
I lolm.m, G. M., B. J. Cook, and R. J. Nachman. 1986. Isolation, pri-
mary structure and synthesis of leucomyosuppressin, an insect neu-
ropeptide that inhibits spontaneous contractions of the cockroach
hindgut. Comp. Bioehem. Physiol 85C: 324-333.
kobierski, L. A., B. S. Beltz, B. A. Trimmer, and E. A. Kravitz.
1987. FMRFamide-like peptides of Homarus americanus: distri-
bution, immunocytochemical mapping, and ultrastructural localiza-
tion in terminal varicosities. J. Camp. Neural. 266: 1-15.
krajniak, K. G. 1991. The identification and structure-activity relations
of a cardioactive FMRFamide-related peptide from the blue crab
Callinectes sapidtix. Peptides 12: 1295-1302.
krajniak, k. G., and D. A. Price. 1990. Authentic FMRFamide is pres-
ent in the polychaete Nereis virens. Peptides 11: 75-78.
kravitz, E. A., S. Glusman, R. M. Harris- Warrick, M. S. Livingstone,
T. Schwarz, and M. F. Goy. 1980. Amines and a peptide as neu-
rohormones in lobsters: actions on neuromuscular preparations and
preliminary behavioural studies. J E\p. Biol. 89: 159-175.
kravitz, E. A., L. kobierski, B. A. Trimmer, and M. F. Goy.
1987. Peptide F,: a myoactive lobster peptide related to FMRF-
amide. Sot Xeurosci. Abslr 13: 1257.
Lange, A. B., 1. Orchard, and V. A. TeBrugge. 1991. Evidence for the
involvement of a SchistoFLRF-amide-like peptide in the neural con-
trol of locust oviduct. J. Comp. Physiol. A 168: 383-391.
Lehman, H. k., and D. A. Price. 1987. Localization of FMRFamide-
like peptides in the snail Helix aspersa. J Exp. Biol 131: 37-53.
Li, C., and R. L. Calabrese. 1987. FMRFamide-like substances in the
leech. III. Biochemical characterization and physiological effects. J.
Neurosa. 1: 595-603.
Marder, E., R. L. Calabrese, M. P. Nusbaum, and B. Trimmer.
1987. Distribution and partial characterization of FMRFamide-like
peptides in the stomatogastric nervous systems of the rock crab. Can-
cer borealis. and the spiny lobster, Paiui/iriis interruptus. J Comp.
Neural. 259: 150-163.
Maynard, D. M. 1960. Circulation and heart function. Pp. 161-226
in Physiology <>l Cnisiticcti. Vol. 1. T. H. Waterman, ed. Academic
Press. New York.
McMahon, B. R., and L. E. Burnett. 1990. The crustacean open cir-
culatory system: a reexamination. Physiol. Zool. 63: 35-71.
McMahon, B. R., I). G. McDonald, and C. M. Wood. 1979. Ventilation,
oxygen uptake, and haemolymph oxygen transport following enforced
340
A. J. MERCIER AND R. T. RUSSENES
exhaustive activity in the Dungeness Crab Cancer nu/xister. J Exp.
Biol. 80:271-285.
McMahon, B. R., and J. L. Wilkens. 1975. Respiratory and circulator)'
responses to hypoxia in the lobster Hamams amcricann\. J. Exp.
Bio/- 62: 637-655.
McMahon, B. R., and J. L. VVilkens. 1977. Periodic respiratory and
circulatory performance in the red rock crab Cancer productus. J
E\P /.ool. 202: 363-374.
Mercier, A. J., I. Orchard, M. Skerrett, and V. TeBrugge.
I991a. FMRFamide-related peptides from crayfish pericardial or-
gans. Soc. Neitrosci. Abstr. 17: 200.
Mercier, A. J., I. Orchard, and V. TeBrugge. 1991 b. FMRFamide-
like immunoreactivity in the crayfish nervous system. J. Exp. Biol.
156: 519-538.
Mercier, A. J., M. Schiebe, and II. I,. Atwood. 1990. Pericardial pep-
tides enhance synaptic transmission and tension in phasic extensor
muscles of crayfish. Neurosci. Lett. 111:92-98.
Painter, S. D., and M. J. Greenberg. 1982. A survey of the responses
of bivalve hearts to the molluscan neuropeptide FMRFamide and to
5-hydroxytryptamine. Biol. Bull 162: 31 1-332.
Price, D. A., and M. J. Greenberg. 1977. Purification and character-
ization of a cardioexcitatory neuropeptide from the central ganglia
of a bivalve molluscs. Prep. Bioc/iem. 7: 261-281.
Price, D. A., and M. J. Greenberg. 1989. The hunting of the FaRPs:
the distribution of FMRFamide-related peptides. Biol. Bull 177: 198-
205.
Price, D. A., W. Lesser, T. D. Lee, K. E. Doble, and M. J. Greenberg.
1990. Seven FMRFamide-related and two SCP-related cardioactive
peptides from Helix. J. Exp. Biol. 154: 421-437.
Robb, S., L. C. Packman, and P. D. Evans. 1989. Isolation, primary
structure and bioactivity of SchistoFLRF-amide, a FMRF-amide-
like neuropeptide from the locust. Scliislocerca gregaria. Bioekem.
Biophys. Res Comni 160: 850-856.
Taylor, A. C. 1976. The respiratory responses of Careinus inaenas to
declining oxygen tension. J. Exp. Biol. 65: 309-322.
Taylor, E. \V., and P. J. Butler. 1978. Aquatic and aerial respiration
in the shore crab Careinus maenas (L.), acclimated to 15°C. J. Comp.
PhvsioL 127: 315-323.
Trimmer, B. A., L. A. Kobierski, and E. A. Kravitz. 1987. Purification
and characterization of FMRFamide-like immunoreactive substances
from the lobster nervous system: isolation and sequence analysis of
two closely related peptides. J. Comp. Neurol. 266: 16-26.
van Harreveld, A. 1936. A physiological solution for freshwater crus-
taceans. Proc. Soc Exp Biol. Med. 34: 428-432.
Watson, \V. H., J. R. Groome, B. M. Chronwall, J. Bishop, and T. L.
O'Donohue. 1984. Presence and distribution of immunoreactive
and bioactive FMRFamide-like peptides in the nervous system of
the horseshoe crab. Limuliis polyphemus. Peptides 5: 585-592.
Watson. W. H., Ill, T. Hoshi, J. Colburne, and G. J. Augustine.
1985. Neurohormonal modulation of the Limuliis heart: amine
actions on neuromuscular transmission and cardiac muscle. J. Exp.
Biol 118: 71-84.
Reference: Biol Bull 182: 341-347. (June. 1992)
Quantitative Analysis by Reverse Phase High
Performance Liquid Chromatography of
5-Hydroxytryptamine in the Central Nervous System
of the Red Swamp Crayfish, Procambams clarkii
GUNDERAO K. KULKARNI AND MILTON FINGERMAN
Department of Ecology. Evolution, and Organismal Biology,
Tulane University, New Orleans, Louisiana 701 18
Abstract. The concentrations of 5-hydroxytryptamine
(5-HT) in central nervous organs of the red swamp cray-
fish, Procambams clarkii, were determined by reverse
phase high performance liquid Chromatography (RP-
HPLC) with electrochemical detection. The quantity
ranged between 54 and 168 pg/mg wet weight of tissue.
The amount is highest in the brain, followed in decreasing
order by the thoracic ganglia, subesophageal ganglion,
eyestalks, and abdominal nerve cord. Significant increases
in the levels of 5-HT in the eyestalks, brain, subesophageal
ganglion, and thoracic ganglia occurred in crayfish ex-
posed for three days to continuous light on a white back-
ground, whereas the 5-HT levels in these tissues decreased
in crayfish kept in darkness. Electrical stimulation of cen-
tral nervous organs in vitro produced significant decreases
in the levels of 5-HT. Fenfluramine (5-HT releaser), 5,6-
DHT (5-HT neurotoxin), and reserpine (5-HT depletor)
induced significant decreases in the 5-HT levels in the
portions of the central nervous system tested.
Introduction
The biogenic amines, norepinephrine. dopamine, his-
tamine. octopamine, 5-hydroxytryptamine (5-HT), and
gamma aminobutyric acid, function as neurotransmitters
in various animals (Werman, 1966; Gerschenfeld, 1973;
Krnjevic, 1974; Fingerman, 1985), and have been found
in crustacean central nervous organs (Beltz and Kravitz,
1983; Elofsson, 1983: Laxmyr, 1984; Fingerman, 1985;
Sandeman el al. 1988). Arechiga el al. (1990) showed
that the species used in this study, the red swamp crayfish
Received 21 October 1 99 1 : accepted 26 March 1992.
Procambams clarkii, contains, in the lamina ganglionaris
of its eyestalks, a set of axons with 5-HT-like immuno-
reactivity. These investigators also found that the respon-
siveness of the retinal photoreceptors of this crayfish to
light is enhanced by exposure to 5-HT, both //; vivo and
in vitro. With respect to other crayfishes, 5-HT-containing
neurons have been reported in the optic lobes and
proto-, deuto-, and tritocerebral regions of the brain of
Pacifastacus leniusculiis ( Myhrberg et al.. 1979; Elofsson,
1983), Orconectes virilis (Sandeman and Sandeman,
1987), and Cherax destructor (Sandeman et al.. 1988).
Kulkarni et al. ( 199 1 ) have shown that 5-HT stimulates
oocyte maturation in Procambams clarkii. This obser-
vation and the earlier study of Arechiga et al. (1990) have
led us to determine, for the first time, the quantity of 5-
HT in the central nervous system of this crayfish. In ad-
dition, we have determined the effects, on the 5-HT con-
centration in components of the central nervous system
of (1) continuous exposure to light or darkness, (2) in
vitro electrical stimulation, and (3) pharmacological agents
known to affect 5-HT levels in vertebrates.
Materials and Methods
Animals
Red swamp crayfish, Procambams clarkii, were pur-
chased from a local seafood dealer and maintained in the
laboratory at 24°C in a recirculating freshwater system.
They were acclimatized to the laboratory conditions
(12:12 L:D) for at least two days before being used in an
experiment. Medium sized (carapace length 40-50 mm),
intermolt (Stage C4, Reddy et al.. 1990) crayfish of both
341
342
G. K. KULKARNI AND M. FINGERMAN
sexes were used. The crayfish were fed commercial crayfish
food. Fiddler crabs, Uca piigilator. were obtained from
the Gulf Specimen Co., Panacea, Florida, and acclima-
tized for three days to the laboratory conditions under
12:12 L:D in a recirculating artificial seawater system.
Tissue preparation and homogenization for 5-HT
del 'emu nations
In this section and the following two, we describe the
procedure for determining the 5-HT concentrations in
the components of the central nervous system of crayfish
maintained in the stock tanks. The eyestalks, brain, sub-
esophageal ganglion, thoracic ganglia, and abdominal
nerve cord of 100 crayfish were dissected out as rapidly
as possible in cold Van Harreveld's crayfish physiological
saline (van Harreveld, 1936). These components were then
distributed, ten per tube, in tubes containing 1 ml of 0.4
M perchloric acid and sonicated in four cycles of 30 s
each with a sonicator (Biosonik-II. Bronwill Scientific)
equipped with narrow probe for small volumes. The ho-
mogenates were centrifuged (10,000 X g) for 15 min at
4°C. The pH of the supernatant, after decanting, was ad-
justed to 6.0 with 2 M potassium carbonate, and the mix-
ture was again centrifuged for 10 min; the supernatants
were used for further purification. The wet weight of each
tissue was recorded. The averages for the ten determi-
nations of the 5-HT contents of each nervous system
component were then calculated.
Purification and analysis of the sample
The supernatants were filtered and purified on a weakly
acid cation exchanger column. Amberlite IRP-64 (Hans-
son and Rosengren. 1978). The column was glass, 30 mm
long with 5 mm i.d.. equipped with a Millipore HV 0.45
nm filter (Nihon Millipore Kogyokk) at the bottom and
filled with 300 /ul of the resin. The supernatant from each
1 ml extract was divided into two 500 ^1 aliquots, and
each aliquot was loaded separately onto the resin at a rate
of 50 ^1/min. The 5-HT adsorbed onto the resin from
each aliquot was eluted with 500 n\ of 1.2 A' HC1. The
two eluted fractions were combined for faster analysis of
only one sample rather than two, and we obtained high
recovery which was always in the range of 80-82%. The
data presented have been corrected to reflect the recovery
percentage. The remaining samples were stored at — 70°C
(Hansson and Rosengren. 1978; Elofsson. ct ai. 1982).
A Waters RP-HPLC unit. Model 501, fitted with a U6K
universal LC injector and a 3.9 X 150 mm 5 ^m silica C-
18 steel column with a small guard column, coupled to
a Waters electrochemical detector (Model 460) was used
for the quantitative analysis of 5-HT in the samples. Five
aliquots (25 /ul) of each sample were run. and the results
were averaged. The variation among samples was less than
10%. The elution reagent was methane sulphonic acid (40
mA/) and phosphoric acid (30 mM} in 17% methanol,
pH 2.5, and was thoroughly degassed before use (Hansson
and Rosengren. 1978; Elofsson el ai, 1982; Nassel and
Laxmyr, 1983). The pressure applied was 1500 psi with
a flow rate of 1 ml/min. The detector was a glassy carbon
electrode, and the working potential was set at +0.75 V
against the reference electrode. The 5-HT concentrations
were determined by comparing the peak height in the
elution profile of the sample with that of the standards,
and are presented as pg/mg of wet tissue. One additional
criterion, other than elution time, was used to identify
the 5-HT peak in the samples. Before analysis, we added
a small amount (10 /ug/ml) of synthetic 5-HT to the sam-
ples. In no case did the sample peak show any inhomo-
geneity due to the addition of 5-HT to the biological ma-
terial when compared with the peak of the standard. With
this analytical system we could detect as little as 25 pg
of 5-HT.
Calibration curve
The calibration curve was prepared as follows. To nine
clean glass tubes, each containing 2 ml of 0.4 M perchloric
acid, was added a known amount of 5-HT creatinine sul-
fate monohydrate (0-51.2 ng/ml free base) and 10~5 M
3,4-dihydroxybenzylamine hydrobromide (DHBA) (in-
ternal standard). The samples were loaded onto the ion
exchange column (Amberlite IRP-64) and treated as
above. Samples ( 100 /^l) were collected and 25 /ul of the
eluate was injected onto the RP-HPLC column. The peak
heights were recorded and fitted in a graph against the
concentration of 5-HT free base. The retention time, with
1 ml/min flow rate and at 1 500 psi. was 8.8-8.9 min for
5-HT and 3.8 min for DHBA (Fig. 1).
Experimental protocols
The 5-HT content in the central nervous organs of the
crayfish were initially determined using specimens that
had been exposed for 2 days to 12:12: L:D. In addition,
20 equal-sized crayfish were held continuously for an ad-
ditional three days either under a fluorescent light (450-
500 lux) or in darkness. After these three days, all of the
central nervous organs from ten crayfish were dissected
out for 5-HT determinations. In addition, the central ner-
vous organs were removed from the rest of these crayfish
and homogenized in 2 ml of crab physiological saline
(Cooke ct ai. 1977). pH 7.4. These extracts were then
centrifuged ( 10,000 X g) at 4°C. and the supernates were
bioassayed for red pigment-dispersing activity in eye-
stalkless fiddler crabs, L'cu piigilator. of 10-15 mm car-
apace width. The pigment in the erythrophores of these
eyestalkless crabs was initially maximally concentrated.
The erythrophores were staged according to the method
5-HYDROXYTRYPTAMINE IN P. CLARKII
343
02468 10 02468 10
MINUTES
Figure 1. Chromatograms of (A) brain supernatant of Praca mbarus
clarkii and (B) aqueous standard containing 5-HT and DHBA. Column:
5 ^m silica C-18 steel. Mobile phase: methane sulphonic acid and phos-
phoric acid in methanol. Detector: glassy carbon electrode at +0.75 V
potential. Retention time for 5-HT 8.8-8.9 min and for DHBA 3.8 min.
of Hogben and Slome (1931) wherein stage 1 indicates
maximal pigment concentration, stage 5 maximal pig-
ment dispersion, and stages 2, 3, and 4 the intermediate
conditions. The Hogben and Slome stages for the exper-
imental and control animals were then used to calculate
Standard Integrated Responses (SIR) of the erythrophores
of Uca pugilator to the nervous tissue extracts ofProcam-
barus clarkii according to the method of Fingerman el
al. (1967). Briefly, when pigment dispersion occurs, the
sum of the Hogben and Slome stages recorded for the
duration of the experiment for the control group is sub-
stracted from the corresponding sum for the experimental
group. The difference is the SIR. The SIR integrates the
amplitude, which is based on the observed Hogben and
Slome stages, and duration of the response of the ery-
throphores. A dose of 50 n\ containing the tissue extract
equivalent to either one eyestalk. brain, subesophageal,
or thoracic ganglion was injected into each crab.
For experiments involving in vitro electrical stimula-
tion, the entire optic tract, including the major ganglia
and sinus gland, was removed from the eyestalk and
maintained in 50 ^1 physiological saline. The remaining
nervous tissue from the brain to the end of thoracic nerve
cord was also dissected out intact and carefully placed in
100 /jl of physiological saline. Electrical stimulation of
the isolated tissue was performed as described in detail
by Quackenbush and Fingerman ( 1984). Briefly, the eye-
stalk tissue was held in place with a suction electrode at-
tached to the stump of the optic nerve, whereas the central
nerve tract, from the brain to the end of the thoracic nerve
cord, was held in place by a suction electrode attached at
the brain end. The stimulation given via the suction elec-
trode was 5 pps. 4 ms delay, with pulses of 40 ms duration
and varying voltage (10. 15. 20. and 25 V). Stimulation
was delivered by a stimulator (Model S44) with a stimulus
isolation unit (Model SIU 5A; both from the Grass In-
strument Co.) After a stimulation bout of 2 min, the ner-
vous tissues were homogenized in 0.4 M perchloric acid
and processed for 5-HT determination as described above.
A total of 75 crayfish were used, divided equally among
one group of unstimulated controls and four groups of
voltage stimulated preparations.
Experiments on the effects of pharmacological agents,
such as fenfluramine (5-HT releaser, Consolo et a/..
1979), fluoxetine (5-HT potentiator, Wong et al.. 1975),
5,6-dihydroxytryptamine (5,6-DHT) (5-HT neurotoxin,
Baumgarten et al., 1982), and reserpine (5-HT depletor,
Myhrberg eta/., 1979: Elofsson et al., 1982), on the levels
of 5-HT in the central nervous tissues of Procambarus
clarkii, were performed according to the procedure of
Myhrberg et al. (1979). Crayfish were divided into 5 groups
of 15 each. The first group received physiological saline
alone and served as the control. The crayfish in the second
through fifth groups were administered various concen-
trations ( 10-25 /ug/g body weight) of either fenfluramine.
fluoxetine. 5.6-DHT, or reserpine. All injections were
given once in a dose of 50 ^1, and the crayfish were sac-
rificed after 2 h and their nervous tissues removed and
processed for 5-HT analysis as described earlier.
Fenfluramine hydrochloride, 5.6-DHT. reserpine. and
Amberlite IRP-64 were purchased from Sigma. Fluoxetine
hydrochloride was a gift from Lilly Research Laboratories,
whereas the 5-HT creatinine sulfate monohydrate and
DHBA were purchased from Aldrich. All drugs were dis-
solved in isosmotic crayfish physiological saline.
The data obtained from these experiments were ana-
lyzed statistically by calculating the standard error for each
of the means (SEM).
Results
The measurements quantifying 5-HT in the central
nervous organs, eyestalks, brain, subesophageal ganglion.
thoracic ganglia, and abdominal nerve cord of crayfish
maintained under laboratory conditions (12:12 L:D;
24nC) in recirculating freshwater for two days are sum-
marized in Figure 2. The 5-HT concentration was highest
in the brain (168 pg/mg) and lowest in the abdominal
nerve cord (54 pg/mg). with the eyestalk, subesophageal
ganglion, and thoracic ganglia having intermediate con-
centrations. Because the concentration of 5-HT in the
abdominal nerve cord is small relative to the rest of the
central nervous organs, only eyestalks. brains, subesoph-
ageal ganglia, and thoracic ganglia were used in the rest
of the experiments.
The pigment in the erythrophores of the crayfish (P.
clarkii) that were illuminated while on a white background
344
G. K. KULKARNI AND M. F1NGERMAN
170
1 50
c?
£
130
110
90
70
50
ES
BR SG
TISSUE
ThG ANC
Figure 2. 5-HT concentration in pg/mg of tissue, assayed by HPLC
and electrochemical detection, in eyestalks (ES), brain (BR), subesoph-
ageal ganglion (SG). thoracic ganglia (ThG). and abdominal nerve cord
(ANC) of Procambarus clarkii. Error bars are SEM for ten separate ex-
tracts of tissue pooled from ten animals each.
was concentrated, whereas the red pigment of the crayfish
kept in darkness was dispersed. The 5-HT concentrations
in the eyestalks, brain, subesophageal ganglion, and tho-
racic ganglia of crayfish held under continuous light for
three days on a white background increased significantly
(Fig. 3). The corresponding red pigment-dispersing SIR
values evoked in the fiddler crabs by these extracts also
increased. In contrast, both the 5-HT levels and the SIR
values in the eyestalks, brains, subesophageal ganglia,
and thoracic ganglia of crayfish held in darkness showed
significant decreases when compared to the controls
(Fig. 3).
In the experiments in which eyestalk, neural ganglia,
and the central nerve tract (from brain to the end of the
thoracic nerve cord) were electrically stimulated with var-
ious voltages (10. 15, 20, and 25 V). 25 V was found to
be most effective. The data in Figure 4 are for tissues
stimulated with 25 V. The stimulation produced signifi-
cant decreases in the concentration of 5-HT in all the
tissues, with the maximum decrease occurring in the
eyestalks ( — 37.4^ ) and the minimum in the thoracic gan-
glia (—22.8%). Furthermore, when extracts of the electri-
cally stimulated tissues were bioassayed for red pigment-
dispersing activity in crabs, it was found that the SIR val-
ues evoked by these extracts were significantly decreased
in comparison to the tissue extracts from the control cray-
fish. Interestingly, the percentage decrease in 5-HT content
decreased progressively in the tissues along the central
nervous chain from the brain to the thoracic ganglia.
Of the concentrations tested, the smallest concentra-
tions that produced significant effects on the 5-HT level
after 2 h were 15 ^g/g body weight of fenfluramine, 10
AJg/g body weight of 5,6-DHT. and 15 yug/g body weight
of reserpine. Fenfluramine induced a significant decrease
of 5-HT from all of the nervous tissues (Fig. 5). The de-
crease was maximum in the eyestalks and least in the
thoracic ganglia. None of the concentrations (10-25 ^g/
g body weight) of fluoxetine produced any significant effect
on the 5-HT concentration in any of the tissues. 5,6-DHT
and reserpine produced significant decreases in the 5-HT
concentration of all the nervous tissues tested. The max-
imum decrease was produced in the eyestalks, whereas
the minimum decrease occurred in the thoracic ganglia.
Discussion
Histochemical studies by means of fluorescence mi-
croscopy have revealed the presence of yellow-fluorescing
Q
205
?
_i
17
185
o
,A
165
"5>
"u
0 ^
13
\
-j ^
Cn « / r
Q. H5
X ^
1
(J
12
J_
L_
T2
z
L?>
12
o
- 125
i-ri
rsj
cn
[<
X
I
•
^r.
Q
i_n
3
9
t>.
9
105
ui
"6"
7
85
r-T
r^
rsj
65
i
C LA DA C LA DA C LA DA
C LA DA
ES BR SG
ThG
TISSUE
Figure 3. 5-HT concentration in eyestalks (ES). brain (BR), sub-
esophageal ganglion (SG) and thoracic ganglia (ThG) of Procambarus
clarkti exposed for three days to continuous light (light adapted. LA) or
held in total darkness (dark adapted, DA). Error bars are SEM often
separate extracts of tissues pooled from ten animals each. Figures in
parentheses are percent change from the 1 2: 1 2 L:D control (C) and those
in the columns are the red pigment-dispersing Standard Integrated Re-
sponses (SIR) of the erythrophores of eyestalkless fiddler crabs. Uca pug-
ilutor. to extract of that tissue of Procambarus clurkn.
5-HYDROXYTRYPTAMINE IN P. CLARK/I
345
170
J_
U
150
o>
E 130
rsi
? 1
_L
12
CL
|
10
oo
(J
10
rsi
§ 110
|
.1".
0
1 —
9
rsi
CO
8
I
rsi
tf> qn
i
J VJ
—
7
m
70
fl
50
C ST C ST C ST C ST
ES BR SG ThG
TISSUE
Figure 4. Effect of 25 V electrical stimulation (ST) for 2 min on the
concentration of 5-HT in eyestalks (ES), brain (BR). subesophageal gan-
glion (SG). and thoracic ganglia (ThG) of Procambarus clarkii. Error
bars are SEM often separate extracts of tissues pooled from ten animals
each. Figures in parentheses denote percent change from the control (C)
and those in columns are the red pigment-dispersing Standard Integrated
Responses (SIR) of the erythrophores of eyestalkless fiddler crabs, i'ca
pugilator. to extracts of that particular tissue of Procambarus clarkii.
cells, indicative of 5-HT, in the eyestalks. brain and ventral
nerve cord of the crayfishes, Astacus astacus (Elofsson et
al, 1966) and Pacifastacus leniusculus (Myhrberg et al.,
1979). The identification of biogenic amines and the de-
termination of their concentrations in arthropods had
earlier depended on fluorometric methods. Those meth-
ods required relatively large tissue samples and yielded
readings only in the microgram or nanogram range. The
advent of HPLC technology enabled investigators to an-
alyze the 5-HT content of tissues from small arthropods
with only moderate amounts of biogenic amines, and the
sensitivity has been extended down to the picogram level.
A comparison of the amount of 5-HT present in various
tissues of the central nervous system of the red swamp
crayfish. Procambarus clarkii, revealed distinct differences
among them (Fig. 2). These experiments provide, for the
first time, data on the amount of 5-HT in different nervous
tissues of this crayfish. The level of 5-HT found in the
eyestalks (102 pg/mg) of Procambarus clarkii is compa-
rable to the value reported by Elofsson et al. (1982) for
the eyestalks (100 pg/mg) of another crayfish, Pacifastacus
leniusciilus, whereas the brain (168 pg/mg) of Procam-
barus clarkii contained slightly more 5-HT than the brain
(150 pg/mg) of Pacifastacus leniusculus (Elofsson et al..
1982). but like Pacifastacus the 5-HT level in the brain
of Procambarus is higher than in the eyestalk.
The 5-HT levels in all nervous tissues of crayfish held
for three days under constant illumination on a white
background were higher than the corresponding values
for the crayfish held for three days in complete darkness,
and the values of the control crayfish held under 12:12
L:D (Fig. 3). Furthermore, the red pigment-dispersing SIR
values for the erythrophores of the eyestalkless fiddler
crabs, Uca pugilator. that received the extracts of eyestalks.
brain, subesophageal ganglion, and thoracic ganglia of the
light adapted crayfish were also significantly higher than
the corresponding SIR values evoked by the nervous tis-
sues of the dark adapted crayfish or by those of the con-
trols. Earlier studies with the fiddler crab, Uca pugilator.
and dwarf crayfish, Cambarellus shufeldti, had revealed
that 5-HT functions as a neurotransmitter that stimulates
the release of red pigment-dispersing hormone (Rao and
Fingerman, 1970, 1975). The changes that occurred in
the 5-HT concentrations in the central nervous organs of
the crayfish in darkness or in constant illumination on a
white background presumably reflect this role of 5-HT in
releasing the color change hormone. The red pigment of
the crayfish kept in light on a white background was con-
centrated. The central nervous tissues of these crayfish on
CO
170
,
-i.
JL
f HO
-v
s
_L
01
CL
2
fT
pi
z 110
o
^
i "
•±1
J-
00
IN
1
r*4
<D
ft
h^
^ 80
i
^
In
in
_JL
<S
50
20
_j_
06
*l
06
irt
1
C FNFL5.6-RS C FNFL5.6-RS
DHT DHT
ES BR
C FNFL5.eRS CFNFL5*-RS
OHT DHT
SG ThG
TREATMENT
Figure 5. Effects of pharmacological agents on the concentration of
5-HT in eyestalks (ES), brain (BR). subesophageal ganglion (SG), and
thoracic ganglia (ThG) of Procambarus clarkii. Error bars are SEM of
ten separate extracts of tissues pooled from ten animals each. Values in
parentheses show the percent change from the control (C). 5.6-DHT
= 5,6-dihydroxytryptamine. 10 ^g/g body weight; FN = fenfluramine.
15 Mg/g body weight; FL = fluoxetine. 25 ^g/g body weight; RS = re-
serpine, 15 ^g/g body weight. The volume of drug solution or saline
injected was 50 ^1. The tissues were removed from the crayfish 2 h after
the injections were given.
346
G. K. KULKARNI AND M. FINGERMAN
a white background not only contained higher 5-HT levels
than the tissues of the control crayfish, but also evoked
higher red pigment-dispersing SIR values, observations
that are consistent with a pigment-dispersing hormone
releasing role of 5-HT. Presumably, because the red pig-
ment was concentrated, neither 5-HT nor pigment-dis-
persing hormone was being used, thereby accounting for
the increased levels of both substances. Earlier, Fingerman
et al. (1964) reported the presence of erythrophorotropic
hormones in the eyestalks and brain of juveniles and
adults of Pwcambarm c/arkii. They observed that the in-
jection of an extract containing one-third of an organ
complement per dose significantly evoked pigment mi-
gration in eyestalkless animals. More recently, McCallum
et al. (1988, 1989) confirmed the previous findings of Fin-
german el al. (1964) by isolating and sequencing the pig-
ment-dispersing hormone (PDH). which is an octadeca-
peptide, from the eyestalks of Procambams clarkii.
Rao and Fingerman (1975) later reported that 5-HT,
when injected into the dwarf crayfish Camharellus s/ni-
feldti, dispersed the red pigment in the erythrophores, as
in the fiddler crab, but was ineffective when tested in vitro
on isolated chromatophore-bearing pieces of the crayfish
carapace. In crayfish on a white background with their
red pigment concentrated, 5-HT turnover would presum-
ably have decreased because, with the red pigment con-
centrated, 5-HT would not be used to stimulate release
of red pigment-dispersing hormone, so a rise in the intra-
neuronal concentration of this neurotransmitter would
occur. On the other hand, because darkness fosters red
pigment dispersion, crayfish in darkness would be using
the intraneuronal stores of 5-HT to effect red pigment
dispersion and would, according to the hypothesis, have
a lower intraneuronal concentration of 5-HT than crayfish
on a white background under light, which the present
data show is indeed the case.
Previously, Berlind and Cooke (1970) reported the re-
lease of a neurosecretory peptide hormone from the peri-
cardia! organs of the spider crabs Libinia emarginata and
Libiniu duhia following electrical stimulation. Later,
Quackenbush and Fingerman (1984) found that electrical
stimulation of the isolated eyestalks of the fiddler crab,
Vca pugilutor. releases chromatophorotropic peptides
from the sinus gland. Recently, Kulkarni and Fingerman
( 1 99 1 ) also used Vca pugilator to show that the distal
retinal pigment light-adapting hormone is released by
electrical stimulation of isolated eyestalk neuroendocrine
tissues. The data presented in Figure 4 clearly show that
25 V stimulation reduced the 5-HT levels in the central
nervous tissues of Procanibarus clarkii. Furthermore, the
stimulation appears also to have reduced the amount of
stored red pigment-dispersing hormone in the central
nervous system because the extracts of the stimulated
tissues produced lesser SIR values for red pigment dis-
persion than did the tissues of the unstimulated controls.
In their studies, both Quackenbush and Fingerman (1984)
and Kulkarni and Fingerman ( 1 99 1 ) bioassayed only the
bathing fluid and not the actual stimulated tissue. In the
present study, the stimulated tissues were extracted and
bioassayed for red pigment-dispersing activity by injecting
the extracts into eyestalkless crabs, Vca pugilator.
Fenfluramine (5-HT releaser), 5,6-DHT (5-HT neu-
rotoxin), and reserpine (5-HT depletor) decreased the
amount of 5-HT in the central nervous system, although
their modes of action are different (Fig. 5). However, the
5-HT potentiator fluoxetine had no appreciable effect on
the 5-HT concentration. These findings are consistent with
earlier 5-HT depletion studies in which reserpine was used
with crustaceans. Myhrberg el ul. (1979), using the his-
tochemical fluorescence method of Falck and Hillarp, and
Elofsson et al. (1982). using HPLC, both found that re-
serpine decreases the 5-HT content of nervous tissues in
the crayfish, Pacifastacus leniiisenhis. Likewise, the 5-HT
content of the brain and eyestalks of Vca pugilator de-
creased after injection of reserpine and 5,6-DHT (Fin-
german et al.. 1974).
Acknowledgments
This research was supported by Grant No. 1-1435-88
from BARD. The United States-Israel Binational Agri-
cultural Research & Development Fund. We thank Mr.
Chayan Chakraborti for his technical assistance.
Literature Cited
Arechiga, H., E. Banuelos, E. Frixione, A. Picones, and L. Rodriguez-
Sosa. 1990. Modulation of crayfish retinal sensitivity by 5-hy-
droxytryptamine. J Exp Bin/. 150: 123-143.
Baumgarten, J. G., S. Jenner, A. Bjorklund, M. P. Klemm, and H. G.
Schlossberger. 1982. Serotonin neurotoxins. Pp. 249-277 in Bi-
ologvol Seriiionerxic Transmission, N. N. Osborne. ed. John Wiley
& Sons, New York.
Beltz, B. S., and E. A. Kravitz. 1983. Mapping of serotonin-like
immunoreactivity in the lobster nervous system. J. Neumsci. 3:
585-602.
Berlind. A., and I. M. Cooke. 1970. Release of neurosecretory hormone
as peptide by electrical stimulation of crab pencardial organs. J. Exp.
Biol 53: 679-686.
Console, S., M. Ladinsky, A. S. Tirelli, V. Crunelli, S. Samanian, and
S. Garattini. 1979. Increase in rat striatal acetylcholine content by
d-fenfluramine, a serotonin releaser. Lite Sci. 25: 1975-1981.
Cooke, I. M., B. A. Haylett. and T. M. Weatherby. 1977. Electrically
elicited neurosecretory and electrical responses of the isolated crab
sinus gland in normal and reduced calcium salines. J. Exp. Biol 70:
125-149.
Elofsson, R. 1983. 5-HT-like immunoreactivity in the central nervous
system of the crayfish. Pacifaslaeits Icnnisculiis. Cell Tiss. Res 232:
221-236.
Elofsson, R., I. Kauri, S. O. Nielsen, and J. O. Slromberg.
1966. Localization of monoaminergic neurons in the central ner-
vous system of Asuiens asuicus Linne (Crustacea). 7. Zellforsch.
Alikrosko/r Anal 107: 343-360.
5-HYDROXYTRYPTAMINE IN P. CLARKII
347
Elofsson. R., L. Laxmyr, E. Rosengren, and C. Hansson. 1982. Iden-
tification and quantitative measurements of biogenic amines and
DOPA in the central nervous system and haemolymph of the crayfish,
Pacilasiacus leniusculus (Crustacea). Comp. Biochem Physiol. 71C:
191-205.
Finjjerman, M. 1985. The physiology and pharmacology of crustacean
chromatophores. Am. /on/. 25: 233-252.
Fingerman, M., \V. E. Julian, M. A. Spirtes, and R. M. Kostrzewa.
1974. The presence of 5-hydro\ytryptamine in the eyestalks and
brain of the fiddler crab L'ca piigilator. its quantitative modification
by pharmacological agents, and possible role as a neurotransmitter
in controlling the release of red pigment-dispersing hormone. ( 'amp
Gen Pharmac. 5: 299-303.
Fingerman, M.. K. R. Rao, and C. K. Bartell. 1967. A proposed uniform
method of reporting response values for crustacean chromatopho-
rotropins: the standard integrated response. Experientia 23:
962-964.
Fingerman, M., Y. Yamamoto. and C. \V. Jacob. 1964. Differences
between the chromatophore systems of juvenile and adult specimens
of the crayfish Procambarus clarkii. Am. Zoo/. 4: 164.
Gerschenfeld. H. M. 1973. Chemical transmission in invertebrate cen-
tral nervous systems and neuromuscular junctions. Physiol. Rc\- 53:
1-119.
Hansson, C., and E. Rosengren. 1978. Quantitative analysis of 5-hy-
droxytryptamine 'n biological material by high performance liquid
chromatography and electrochemical detection. Anal Lett BI1:
901-912.
Hogben, I... and D. Slome. 1931. The pigmentary effector system- VI.
The dual character of endocrine co-ordination in amphibian colour
change. Proc. R Soc. Land. 108B: 10-53.
Krnjcvic, K. 1974. Chemical nature of synaptic transmission in ver-
tebrates. Physiol. Rev. 54: 418-540.
Kulkarni, G. K., and M. Kingerman. 1991. Release of black pigment-
dispersing hormone and distal retinal pigment light-adapting hormone
upon electrical stimulation of the isolated eyestalk neuroendocrine
complex of the fiddler crab. L\ a piigilator. Comp. Biochem Physiol.
99C: 47-52.
kulkarni, G. K.. R. Nagabhushanam, G. Amaldoss, R. G. Jais»al, and
M. Fingerman. 1991. 5-hydroxytryptamine stimulation of the ovary
in the crayfish, Procambarus clarkii. Am. Zoo/. 31: 1 15A.
I.axmyr, L. 1984. Biogenic amines and DOPA in the central nervous
system of decapod crustaceans. Camp. Biachcm Physiol 77C:
139-143.
McCallum, M. L., K. R. Rao, and J. P. Riehm. 1989. A comparison
of the effects of two pigment-dispersing hormones on the crayfish
Procambarus clarkii Am ./no/ 29:217.
McCallum, M. L., K. R. Rao, J. P. Riehm, C. J. Mohrherr, and \V. T.
Morgan. 1988. Isolation of a (5-PDH analog from the crayfish. Pro-
cambarus clarkii. Bio/. Bull 177: 225-229.
Myhrberg, H. E., R. Elofsson, R. Aramant, N. Klemm, and I.. I.axmyr.
1979. Selective uptake of exogenous catecholamines into nerve
fibres in crustaceans. A fluorescence histochemical investigation.
Comp. Biochem Physiol. 62C: 141-150.
Nassel, D. R., and I.. Laxmyr. 1983. Quantitative determination of
biogenic amines and DOPA in the CNS of adult and larval blowflies.
Calliphora erylhraccphtila Comp. Biochem. Physiol. 75C: 259-265.
Quackenbush, L. S., and M. Fingerman. 1984. Regulation of the release
of chromatophorotropic neurohormones from the isolated eyestalk
of the fiddler crab. L 'ca pugilalor. Biol Bull. 166:237-250.
Rao, K. R., and M. Fingerman. 1970. Action of biogenic amines on
crustacean chromatophores. II. Analysis of the responses of ery-
throphores in the fiddler crab, L'ca piigilator. to indolealkylammes
and an eyestalk hormone. Comp. Gen. Pharmacol. 1: 1 17-126.
Rao. K. R., and M. Fingerman. 1975. Action of biogenic amines on
crustacean chromatophores. IV. Analysis of the synergistic erythro-
phoric pigment dispersion evoked by 5-hydroxytryptamine and lys-
ergic acid diethylamide in the dwarf crayfish. Cambarellits smitcldn
Comp. Biochem. Physiol 51C: 53-58.
Reddy. T. S. N., G. K. Kulkarni, I,. Glade, and M. Fingerman.
1990. Molt stages of the red swamp crayfish. Procambarus clarkii.
Am Zoo/. 30: 109A.
Sandeman, R. E., and D. C. Sandeman. 1987. Serotonin-like immu-
noreactivity of giant olfactory interneurons in the crayfish brain. Brain
Res. 403: 371-374.
Sandeman, D. C., R. E. Sandeman, and A. R. Aitken. 1988. Atlas of
serotonin-containing neurons in the optic lobes and brain of the
crayfish. Cherax destructor J. Comp \eurol 269:465-478.
van Harreveld, A. 1936. A physiological solution for fresh water crus-
taceans. Proc. Soc. E.\p Biol Med. 34: 428-432.
Werman, R. 1966. Criteria for identification of a central nervous system
transmitter. Comp. Biochem Physiol. 18: 745-766.
Wong, D. T., F. P. Bymaster, J. S. Horng, and B. B. Molloy. 1975. A
new selective inhibitor for uptake of serotonin into synaptosomes ot
rat brain: 3-(p-trifluoromethylphenoxy)-N-methy!-3-phenylpropyl-
amme. ./ Pharmacol Exp. Thcr 193:804-811.
Reference: Biol. Bull 182: 348-365. (June. 1992)
New Interpretation of a Nudibranch Central
Nervous System Based on Ultrastructural Analysis
of Neurodevelopment in Melibe leonina.
I. Cerebral and Visceral Loop Ganglia
LOUISE R. PAGE
Department of Biology, University of Victoria, Victoria, British Columbia, Canada V8W 2Y2
Abstract. Development of the 'cerebropleural' ganglia
in the dendronotid nudibranch Melibe leonina (Gould.
1 852) was examined by electron microscopy of semi-serial
sections through larval stages. Although comparative
neuroanatomical studies suggest that the paired cerebro-
pleurals of nudibranchs are formed by fusion of cerebral
and pleural ganglia, plus all other ancestral ganglia of the
visceral loop, my study indicates that the pleural ganglia
are not part of these compound ganglionic masses. In
Melibe larvae, the cerebral, optic, and rhinophoral ganglia,
arise from pre-trochal cephalopedal ectoderm. At hatching
stage, the visceral loop extends from the two cerebral gan-
glia, is non-ganglionated, and forms a complete circuit
beneath the esophagus. Ganglia that subsequently develop
along the visceral loop, which were identified as subin-
testinal, visceral, supraintestinal, and possibly right pa-
rietal ganglia, arise from placodes of visceropallial ecto-
derm that show torsional displacements. In addition, a
cluster of neurons, presumed to be osphradial, lies close
to the rim of the right mantle fold. Detorsion of the visceral
loop is accomplished by migration of subintestinal neu-
rons along the visceral loop fiber tract, not by visceral
loop shortening. Localized elongation of a different seg-
ment of this fiber tract during metamorphosis displaces
the visceral ganglion to the left, where it fuses with sub-
intestinal and left cerebral ganglia.
Introduction
More than a century of comparative neuroanatomical
studies on opisthobranchs have revealed a wide range of
ganglionic fusions and cephalization within this gastropod
Received 6 November 1991; accepted 5 March 1992.
group (Russell. 1929; Hoffmann, 1936; others reviewed
by Bullock, 1965; Schmekel, 1985). In species showing
evidence effusions, homologous ganglionic regions have
been inferred by extrapolation from the layout of distinct
ganglia, connectives, and peripheral nerves found in pre-
sumably more primitive species. Although these inter-
pretations have been largely accepted by most contem-
porary gastropod systematists and neurophysiologists, they
are nevertheless conjectural (see Dorsett, 1986). This is
particularly true for nudibranchs, in which ganglia of the
central nervous system (CNS) show extreme consolida-
tion. Ambiguity about homologous ganglionic regions re-
duces the taxonomic value of neuroanatomical characters
and contributes to the uncertainty about phyletic origins
and relationships of the opisthobranchs (see Minichev,
1970: Minichev and Starobogatov. 1978: Gosliner, 1981,
1991; Gosliner and Ghiselin, 1984; Haszprunar. 1985b,
Schmekel. 1985). Furthermore, possible misconceptions
about ganglionic fusion patterns can confound compar-
ative neurobiological studies — a lamentable situation for
an animal group that is otherwise very amenable to neu-
roethological investigation (see Kandel, 1979; Willows,
1985-1986).
Studies of gangliogenesis in prosobranch gastropods
indicate that the CNS typically develops from a similar
groundplan, with various derived conditions arising later
in ontogeny [compare studies of Crofts (1937) and Moritz
(1939) with those of Honegger (1974) and Demian and
Yousif ( 1975)]. This groundplan, in which discrete ganglia
are interconnected in a specific pattern, is not greatly al-
tered during the subsequent development of some pro-
sobranchs. and it appears also in Gosliner's (1981) pro-
posal for the ancestral opisthobranch nervous system.
348
NUDIBRANCH NEUROGENES1S I
349
Figure 1. Opisthobranch central nervous systems showing varying
degrees of cephalization. ganglionic fusions, and euthyneury. Dorsal
views: cerebral ganglia stippled. A. Possible ancestral condition (adapted
from Gosliner. 1981): distinct parietal ganglia found in some extant op-
isthobranchs. B. Aplysiu californica (adapted from Knegstein, 1 979a. b).
C. aeolid nudibranch (adapted from Russell, 1929): note visceral loop
emerging from ganglionic mass projecting posteriorly from cerebral gan-
glia. AG = abdominal ganglion; BG = buccal ganglia; OG = osphradial
ganglion = PG = pedal ganglion; PAG = parietal ganglion; PLG = pleural
ganglion; SBG = subintestinal ganglion; SPG = supraintestinal ganglion;
VG = visceral ganglion: VL = visceral loop.
Reasoning from studies of a variety of opisthobranchs.
Gosliner (1981) suggested that the common ancestor of
this group had paired cerebral, pleural. and pedal ganglia
interconnected around the esophagus, a long streptoneu-
rous visceral loop (twisted due to torsion) punctuated by
subintestinal, supraintestinal, and visceral ganglia, and an
osphradial ganglion connected to the supraintestinal gan-
glion (Fig. 1 A). Extant opisthobranchs also have a pair of
buccal ganglia, and some species have an extra ganglionic
pair, called the parietals, located anterior to the intestinals.
Developmental studies on the anaspidean Aplysia cali-
Jornicu have shown that all ganglia of Gosliner' s ancestral
opisthobranch appear during early larval development,
but the visceral and intestinal ganglia fuse eventually to
form what is called the abdominal ganglion (Fig. IB)
(Knegstein. 1977a. b: Schacher el a/.. 1979a, b). These
results suggest that the true pattern of ganglionic fusions
among other opisthobranchs might be revealed by studies
of neurodevelopment.
The adult CNS of nudibranchs. which shows only three
pairs of distinct ganglia surrounding the esophagus, is
much more consolidated than that of aplysiids. The gan-
glia have been identified traditionally as the pedals, buc-
cals. and cerebropleurals. with the latter incorporating the
cerebral and pleural ganglia plus all other ganglia of the
visceral loop (Fig. 1C). Developmental studies of nudi-
branchs have indeed shown that the cerebropleurals are
constructed ontogenetically from precursor ganglia, but
there are three different interpretations for the location
and identity of these precursor ganglia (Thompson, 1 958;
Tardy, 1970, 1974; Bickell and Chia, 1979; Bickell and
Kempf. 1983: Kempf iVrt/., 1987). Much of this confusion
may stem from limited resolution provided by histological
sections.
To address the controversy regarding the ontogeny of
the cerebral and visceral loop ganglia in nudibranchs, I
have cut semi-serial, ultrathin sections through sequential
larval stages of the dendronotid nudibranch Melihe leo-
nina. A review of the genus has been given recently by
Gosliner (1987). General features of larval and meta-
morphic development in this species were described from
histological sections by Bickell (now Page) and Kempf
( 1983). The interpretation of gangliogenesis given in the
present paper and the following companion paper (Page.
1992) differs from that described in the earlier study.
Materials and Methods
Adults of Melibe leonina and their egg masses were
collected from Patricia Bay, Vancouver Island, Canada.
Figure 2. Veliger larvae of Melibe leonina. A. Lateral view, prior to
mantle retraction, showing basic anatomy; gut is stippled. Broken line
passes along floor of mantle cavity and demarcates cephalopeda! mass
from visceropallial mass. Arrow indicates displacement of mantle at
mantle retraction. B. Oblique, antero- ventral view of young veliger; right
velar lobe cut away to reveal mantle fold lining right mantle cavity.
Swellings in mantle fold are osphradial neurons (OS) and apices of mantle
gland (MG). The asterisk marks position of right pallial placode; open
arrow indicates a site beneath the foot, where the left pallial placode is
located. Note invagination of left cephalic plate (CP) within pre-trochal
ectoderm. C. Dorsal view of veliger shortly after mantle retraction showing
positions of cerebral ganglia and components of visceral loop (developing
CNS stippled). Mantle gland omitted for clarity. A = anus; CO = cerebral
ganglion: CP = cephalic plate; EY = eye; E = esophagus; F = foot; I
= intestine; LOG = left digestive gland; LPP = left pallial placode; LRM
= larval retractor muscle; M = mouth; MA = shell-secreting cells of
mantle: MG = mantle gland; NP = nephrocyst pore; O = operculum;
PR = prototroch (cilia not shown in 2B); S = stomach; SBG = subin-
testinal ganglion; SH = shell; SPG = supraintestinal ganglion: ST = sta-
tocyst; V = velar lobe; VP = visceral placode.
350
Figure 3. Summary sketches of developing ganglia of cephalic plate
and visceral loop in Mtiihc lamina; postero-lateral views from right side.
A. newly hatched larva. B. larva at mantle retraction stage. C. late larval
stage. D. metamorphic stage. CC = cerebral commissure; CG = cerebral
ganglion; EY = eye: LPP = left pallial placode; OS = osphradial neurons;
RG = rhinophoral ganglion; SBG = subintestinal ganglion; SPG = su-
praintestinal ganglion; ST = statocyst; VG = visceral ganglion; VL
= visceral loop: VP = visceral placode.
Larvae that hatched from egg masses laid in the field or
laboratory were reared according to the method of Bickell
and Kempf (1983). Under laboratory conditions and a
rearing temperature of 1 2 to 1 4°C, larvae required a min-
imum of 5 weeks to complete pre-metamorphic devel-
opment.
Larvae were anaesthetized as described by Bickell and
Kempf (1983), and were fixed according to the method
of Bickell and Chia( 1979).
The area of the larval body containing developing gan-
glia was thin sectioned in whole or part with a diamond
knife, and batches of eight to ten sections were collected
on uncoated copper grids ( 1 50 mesh size). The grids were
first washed in acetone and distilled water, then passed
briefly through the flame of an alcohol burner; this eases
the pick-up of floating sections, possibly by reducing the
hydrophobicity of the grid surface. To discourage the ten-
dency of sections to float towards the grid periphery, grids
were bent slightly so that sections were lifted onto a convex
grid surface without excess water. The central areas of wet
sections were teased over openings between grid bars with
an eyebrow hair mounted on an orange stick. With this
method, two to six whole sections per grid could be
viewed, with each gridload of sections representing ap-
proximately 0.8 fj.m of tissue thickness. Sections were
stained for 90 min in aqueous 2% uranyl acetate and 8
min in 0.2% lead citrate at room temperature.
Initially, a general picture of the CNS within each spec-
imen was reconstructed by photographing one section per
two to three grids at a magnification of 873x on the neg-
ative with a Philips EM300. This gave a panoramic view
of the section but still allowed resolution of small fiber
tracts. Subsequently, areas showing important structural
features were photographed at higher magnification from
these and intervening grids.
The following larval stages were thin sectioned: newly
hatched, 6 days old, just prior to mantle retraction (larval
shell at full size but mantle fold not yet retracted), onset
of mantle retraction, complete mantle retraction, hyper-
trophy of retracted mantle fold, and late stage larva with
ceratal rudiments. Characteristics of the mantle fold were
used to stage larvae because these can be recognized before
sectioning (Bickell and Kempf, 1983). Stages fixed im-
mediately after shell loss, and at 5 and 10 h after shell
loss, were cut into serial 1 urn thick sections and stained
with methylene blue and Azure II in sodium borax (Rich-
ardson et nl.. I960). All larval stages were also thick sec-
tioned to corroborate gross features seen in ultrathin sec-
tions.
Results
General
Planktotrophic veliger larvae ofMelibe Iconina consist
of two major parts: a cephalopedal mass, which includes
the ciliated velar lobes, distal esophagus, and foot; and a
visceropallial mass, which consists of the remainder of
the gut encased by the mantle (pallium) (Fig. 2A). The
mantle is derived from the embryonic shell gland and
consists of a squamous epithelium lining the inner wall
of the shell and a peripheral rim of large cells that secrete
larval shell material. Epithelium extending from the shell-
secreting cells to the cephalopodium is called the mantle
fold, although it is unclear if this is also derived from the
shell gland. In newly hatched Melibe larvae, the mantle
fold is onlv a few cells wide and thus the mantle cavitv it
Figure 4. Cross section through newly hatched larva passing through cerebral ganglia (CG). cerebral
commissure (CC). and cephalic apical organ (AO). The prototroch (PR) is indented over the mouth. Arrow
indicates position of osphradial nerve (enlarged in inset) within the right mantle fold (MF). Orientation
arrows: D = dorsal: V = ventral; L = left; R = right. Scale, 10 ^m.
Figure 5. Cross section through newly hatched larva showing part of cerebral commissure (CC) extending
into left cerebral ganglion (CG). Arrow indicates cilia within apical organ cell. Scale, 2 nm.
Figure 6. Cross section through apex of 6-day-old larva showing cephalic plates (CP) invaginated from
pre-trochal ectoderm. PR = prototroch. Scale. 10 pm.
Figure 7. Frontal section through left side of larva at onset of mantle retraction showing multilayered,
invaginated cephalic plate (CP) overlying cerebral ganglion (CG). Arrowhead indicates mitotic cell: arrow
indicates ingressing cephalic plate cells. V = velar lobe. Scale. 5 jim.
Figure 8. Frontal section through right side of larva at onset of mantle retraction showing eye (EY)
overlying dorsal area of cerebral ganglion (CG). CC = cerebral commissure. Scale, 5 fim.
NUDIBRANCH NEUROGENES1S I
351
352
L. R. PAGE
defines is extremely shallow. The anus opens into the
mantle cavity on the right ventro-lateral side, indicating
slightly less than 90° torsion of the gut (Fig. 2B).
The shell growth that occurs during the first half of
larval life is accompanied by a deepening of the right side
of the mantle cavity, which extends from the anus over
to the dorsal side (Fig. 2B). The left mantle cavity, which
extends from the anus along the ventral aspect of the larva,
remains shallow. It is important to note that the right
mantle fold in partially torted Melibe larvae is equivalent
to the left mantle fold of gastropods showing full 180° of
torsion. It is this side that retains the various components
of the pallial complex (ctenidium, osphradium, kidney)
in monotocardian prosobranchs.
Midway through the larval phase of Melibe, shell se-
cretion is arrested and the mantle fold detaches from the
shell aperture and retracts posteriorly. The future post-
metamorphic dorsum and cerata are formed from re-
tracted mantle fold.
Two days after metamorphic shell loss, the basic char-
acteristics of the adult nervous system are evident. The
post-metamorphic CNS is formed by consolidation of
many ganglionic primordia that arise during larval de-
velopment by cellular ingression from thickened placodes
of ectoderm. Neurogenic ectodermal cells often show mi-
totic figures.
To identify ganglionic primordia in Melibe, the loca-
tions of their respective neurogenic ectodermal placodes
were compared to those that give rise to specific ganglia
in prosobranchs. For these comparisons, it is important
to distinguish three main areas of larval ectoderm, as il-
lustrated in Figure 2. Ectoderm of the cephalopedal mass
is subdivided into pre- and post-trochal areas by the band
of large velar ciliated cells (the prototroch) that runs
around the periphery of the velar lobes (Fig. 2B). The
distinction between cephalopedal and visceropallial ec-
toderm is less distinct, but the latter shows torsional dis-
placement whereas cephalopedal ectoderm does not. The
pores of the left and right nephrocysts (large terminal cells
of protonephridia) provide a convenient marker for the
posterior limit of cephalopedal ectoderm (Fig. 2B). The
nephrocysts do not exhibit torsional displacement in
young larvae, although the left member is pulled further
posteriorly than the right during mantle retraction. The
anus and mantle gland (previously called larval kidney
cell) lie posterior to the nephrocysts and show evidence
of torsion (Figs. 2A, B).
The development and post-metamorphic fate of ganglia
derived from pre-trochal cephalopedal ectoderm and from
ectodermal placodes within the visceropallium are de-
scribed below. Results are summarized by the diagrams
in Figures 2C and 3. The companion paper (Page, 1992)
describes ganglionic derivatives of post-trochal cephalo-
pedal ectoderm.
Pre-trochal cephalopedal ganglia
Two placodes of neurogenic ectoderm, called cephalic
plates, flank the mid-sagittal plane of the pre-trochal ce-
phalopedal zone (Fig. 2B). Cells ingressing from the ce-
phalic plates form the paired cerebral ganglia, eyes, optic
ganglia, and rhinophoral ganglia.
In newly hatched veligers, each cerebral ganglion con-
sists of neuronal cells clustered around a neuropile that
is continuous with the fiber tract of the cerebral commis-
sure. In cross sections, the cerebral ganglia and commis-
sure form a horseshoe shaped complex (Fig. 4), perched
above the distal esophagus, that parallels the trajectory of
the prototroch where it arches above the mouth. The pe-
culiar cells of the cephalic apical organ, which have an
internal vacuole containing many cilia, are located on the
dorsal side of the cerebral commissure (Fig. 5).
The paired cephalic plate placodes that overlie the ce-
rebral ganglia are invaginated only slightly at hatching,
but 6 days later the invaginations have deepened and are
connected by a trough (Fig. 6). By the time of mantle
retraction, the two invaginated cephalic plate lobes are
composed of multiple layers of cells that sit as caps over-
hanging the ventro-lateral borders of the enlarging cerebral
ganglia (Fig. 7). These cells ingress individually or in small
clusters to the cerebral ganglia. Evidence of ingression is
most prominent along the medial side of each invaginated
cephalic plate lobe (Fig. 7).
Two eyes, each with a spherical lens and pigment gran-
ules, appear on the dorso-lateral surface of each cerebral
ganglion prior to mantle retraction (Fig. 8). The eyes form
adjacent to protuberances of the cerebral ganglia, which
become the optic ganglia. The rhinophoral ganglia, which
also arise from the cephalic plate, appear as mounds of
cells on the cerebral ganglia of late stage larvae. During
Figure 9. Cross section through newly hatched larva at a level just beneath the foot showing where the
visceral loop completes its subesophageal trajectory. Area within left box enlarged in Figure 10; area within
right box enlarged in Figure 1 1. A = anus; E = esophagus; LPP = left pallial placode; LRM = larval retractor
muscle; MA = shell-secreting mantle cells; MG = mantle gland. Scale, 10 ^m.
Figure 10. Left limb of visceral loop (arrows) extending towards right side beneath esophagus (E). Scale,
2 /im.
Figure 11. Right limb of visceral loop (R) leaving mantle gland (MG) to merge with left limb (L). A =
anus. Scale, 2 ^m.
NUDIBRANCH NEUROGENESIS I
353
; - -.
354
L. R. PAGE
metamorphosis, the rhinophoral ganglia are carried dis-
tally within the expanding oral hood (Bickell and Kempf,
1983).
Visceral loop ganglia
To minimize confusion, I will define my nomenclature
for the visceral loop ganglia because varying schemes have
been used by past authors. I use the terminology of Bullock
( 1965), in which the most anterior ganglia of the typical
gastropod visceral loop are called left and right pleurals
(Fig. 1A). These are followed by the sub- and supraintes-
tinals and finally the unpaired visceral ganglion. The su-
praintestinal ganglion gives rise to a nerve extending to
the osphradial ganglion. A second osphradial ganglion
linked to the subintestinal ganglion is a characteristic re-
stricted to diotocardian prosobranchs. In some pulmo-
nates and opisthobranchs, but not in prosobranchs, extra
ganglionic swellings are located anterior to the intestinal
ganglia; these are called parietal ganglia. The visceral loop
is long and twisted (streptoneurous) in many prosobranchs
and a few opisthobranchs, but is short or untwisted (eu-
thyneurous) in most extant opisthobranchs.
Sections through newly hatched larvae of Melibe show-
that the left and right limbs of the visceral loop fiber tract
emerge directly from the neuropiles of their respective
cerebral ganglia. The left limb swings beneath the esoph-
agus to merge with the right limb at a point just medial
to the anus (Figs. 9-11). The visceral loop is complete but
nonganglionated at this initial larval stage and consists of
at least 20 axons.
Figure 1 2 is a cross section through the left side of a
newly hatched veliger, close to where the visceral loop on
that side emerges from the cerebral ganglion. The section
gives the impression that post-trochal ectodermal cells are
ingressing from the left mantle fold toward the base of
the cerebral ganglion. However, higher magnification re-
veals myofilaments within these subepidermal cells (Fig.
13); they are not ingressing pleural or parietal neurons.
Differentiating subepidermal muscle cells are present in
the same location on the right side of the larva. Indeed,
the nuclear regions of many other fully and partially dif-
ferentiated muscle cells are concentrated within the ceph-
alopedal area of young larvae; these cannot be distin-
guished from neuronal elements in histological sections.
Right limb of visceral loop and osphradial neurons
After travelling a short distance posteriorly from the
right cerebral ganglion, the right limb of the visceral loop
in newly hatched veligers associates with a placode of
thickened ectoderm that I call the right pallial placode
(Fig. 14). The right pallial placode lines the deepest part
of the right mantle cavity, immediately beneath the pore
of the ipsilateral nephrocyst (Fig. 2B). At least one neuron
originating from the right pallial placode is fully differ-
entiated at hatching stage and is filled with many vesicles
(Fig. 14). From this point, a small nerve extends anteriorly
between the epithelial cells and underlying basal lamina
of the mantle fold (Figs. 4 inset, 15). At least some of the
axons forming this nerve arise from a cluster of sparsely
ciliated neurons located close to the antero-lateral pe-
riphery of the right mantle fold. These neurons are as-
sociated with a mucous cell and their location corresponds
to that of the osphradium and osphradial ganglion in pro-
sobranch larvae (Figs. 2B, 15, 16). The fate of these os-
phradial neurons after mantle retraction requires further
study. I failed to find them in sections of larvae fixed after
onset of mantle retraction.
At least some of the neurons ingressing from the right
pallial placode must be homologues of supraintestinal
neurons, because they form a ganglion at the intersection
of the visceral loop and osphradial nerve. However, the
series of sections shown in Figures 17-20 reveal that this
ganglion receives neurons from two ingression sites within
the right pallial placode. These two sites may be sources
of right parietal and supraintestinal neurons. However,
an alternative interpretation is given in the discussion,
and until this issue is clarified I will refer to the ganglion
arising from the right pallial placode as the supraintestinal
ganglion.
Figure 12. Cross section through left side of newly hatched larva close to where visceral loop (VL)
emerges from cerebral ganglion (CO). Arrow indicates subepidermal cell (enlarged in Fig. 1 3) lying between
mantle fold ectoderm (MF) and cerebral ganglion. MC = mantle cavity; ST = statocyst. Scale, 2 ^m.
Figure 13. Detail from Figure 12 showing myofilaments (arrows) within subepidermal cell. Scale, 0.5
^m.
Figure 14. Cross section through newly hatched larva showing right limb of visceral loop (VL) deflecting
towards right pallial placode (RPP). Asterisk marks neuron with many vesicles. Scale, 1 /^m.
Figure 15. Mantle fold ectoderm on right side of 6-day-old larva showing differentiating osphradial
neuron (asterisk) adjacent to axons of osphradial nerve (arrowheads). Arrow indicates vesicles that are
enlarged in the inset. Scale, 1 pm; inset, 0.3 ^m.
Figure 16. Frontal section through larva just prior to mantle retraction showing osphradial neurons
(OS) and associated mucous cell (arrow) beneath shell-secreting periphery of mantle (MA). Supraintestinal
ganglion (SPG) enlarged in Figure 19. H = hemocoel: N = nephrocyst; RPP = right pallial placode. Scale.
NUDIBRANCH NEUROGENESIS I
355
MC
PSM1
15
356
L R. PAGE
Immediately after leaving the supraintestinal ganglion,
the right limb of the visceral loop passes to the very large
mantle gland that bulges into the hemocoel from the right
mantle fold ectoderm (Fig. 20). This structure has been
called the larval kidney complex, but evidence of an ex-
cretory role is weak and its ultrastructure is more consis-
tent with a secretory function. The largest cell of the man-
tle gland, along with several associated secretory cells and
the muscle fibers that invest the complex, are innervated
by axons extending from the intimately associated visceral
loop fiber tract.
In young larvae, the right limb of the visceral loop trav-
els ventrally along the mantle gland until merging with
the left limb (Fig. 11). However, the visceral loop lifts
away from the mantle gland when the latter is displaced
posteriorly during mantle retraction (Fig. 2 1 ). Although
the peripheral rim of the mantle fold is pulled a great
distance posteriorly during mantle retraction, the right
pallial placode continues to be closely associated with the
supraintestinal ganglion (Fig. 22). The ganglion lies lateral
to the esophagus but projects dorsally.
Left limb of visceral loop and visceral ganglion
At the hatching stage, mantle ectoderm to the left of
the anus is thickened by presumptive neurons. I call this
the left pallial placode, because it is part of the left mantle
fold. However, due to torsion, the placode is located to-
ward the right side of the ventral aspect of the larval body
(Figs. 2B, 9, 10). During later development, the left pallial
placode is the only recognizable source of neurons along
the entire left limb of the visceral loop in all larval stages
examined. Furthermore, the only axons that leave this
portion of the visceral loop extend to larval muscles, not
to the ectoderm. Nevertheless, during subsequent devel-
opment, cells are distributed along the entire left limb of
visceral loop with a concentration appearing where the
visceral loop emerges from the left cerebral ganglion (Fig.
23). These cells appear to be subintestinal neurons that
originated from the left pallial placode on the right, ventro-
lateral side of the body and that migrated along the sub-
esophageal trajectory of the visceral loop toward the left
cerebral ganglion. Unlike the supraintestinal ganglion,
which projects dorsally from the right cerebral ganglion
(Fig. 22), the concentration of subintestinal neurons pro-
jects ventrally from the left cerebral ganglion (Fig. 23).
I could find no morphological evidence of a distinct
ingression site for left parietal neurons.
The visceral placode, which is neurogenic visceropallial
ectoderm for neurons of the future visceral ganglion, be-
comes recognizable at 6 days after shell loss. The visceral
and left pallial placodes are contiguous. Figures 24-27,
which is a series of frontal sections through a larva at
onset of mantle retraction, show the positions of these
two placodes and their relationship to the visceral loop.
Note that the right limb of the visceral loop extends to
the left pallial placode, not to the visceral placode.
During later development, the left limb of the visceral
loop becomes progressively denuded of cells, presumably
because most subintestinal neurons have migrated toward
the left cerebral ganglion. Nevertheless, a small clump of
neurons is apparent at the junction of the left and right
limbs of the visceral loop and from this point, a prominent
visceral nerve extends into the base of the visceral placode
(Figs. 28-30). Ingressing visceral neurons do not form a
ganglion immediately beneath the visceral placode. My
observations suggest that visceral neurons migrate along
the visceral nerve to form a consolidated ganglion on the
visceral loop.
By the end of the larval phase, there are three neuronal
concentrations along the visceral loop fiber tract: the su-
praintestinal ganglion behind the right cerebral ganglion,
the subintestinal ganglion behind the left cerebral gan-
glion, and the visceral ganglion immediately adjacent to
the anus (Fig. 3C). At this stage, the two intestinals have
begun to fuse with their respective cerebral ganglia, but
the visceral ganglion remains separate. Histological sec-
tions through metamorphic stages show that the visceral
ganglion is relocated to the left side during the hours after
shell loss, where it fuses with the left cerebral and sub-
intestinal ganglionic mass (Figs. 3D; 31-33).
Discussion
It has been known since the last century that molluscan
neurons are ectodermal derivatives (see reviews by Raven.
Figures 17-20. Series of frontal sections through nght pallial placode (RPP) and supraintestinal ganglion
(SPG) of a larva just prior to mantle retraction. N = nephrocyst.
Figure 17. Arrow indicates presumptive neurons ingressing from nght pallial placode to underlying
ganglion. Scale, 2 urn.
Figure 18. Slightly deeper section showing no neuronal ingression. Scale. 2 ^m.
Figure 19. Second site of neuronal ingression (arrowl coinciding with emergence of osphradial nerve
(arrowheads) from the visceral loop (VL). Scale, 2 /im.
Figure 20. Visceral loop (VL) travelling from supraintestinal ganglion to mantle gland (MG). Scale. 2
Figure 21. Longitudinal section through supraintestinal ganglion (SPG) at full mantle retraction stage
showing visceral loop (arrow) dissassociated from mantle gland (MG). Scale. 5 ^m.
NUDIBRANCH NEUROGENES1S I
357
358
L R. PAGE
1958; Moor, 1983). Furthermore, comparison of various
histological accounts of prosobranch neurogenesis indicate
that primordia of each ganglionic type arise from ste-
reotypic locations within the ectoderm of the veliger
(Smith, 1935; Crofts, 1937; Moritz, 1939; Creek, 1951;
Regondaud, 1961, 1964; D'Asaro, 1969; Cumin. 1972;
Guyomarc'H-Cousin. 1974; Honegger. 1974; Demian
and Yousif, 1975). Cerebral ganglia always arise from the
intravelar cephalic plates and intestinal, visceral, and os-
phradial ganglia arise from characteristic areas of viscer-
opallial ectoderm. This phenomenon is particularly un-
ambiguous in those gastropods that retain separate central
ganglia through metamorphosis. Nevertheless, these sites
appear to be a highly conserved feature of early gastropod
neurogenesis, regardless of the final, species-specific form
of the adult CNS (illustrated well by studies of Honegger,
1974; Demian and Yousif, 1975). Therefore, it is appro-
priate to identify ganglionic primordia in Melibe by com-
paring their ectodermal ingression sites to those described
for the ganglia of other gastropods, particularly proso-
branchs. Adjustments for variable degrees of torsional
displacement among these species are made possible by
the position of the anus relative to the cephalopedal mass.
Using in situ hybridization and immunofluorescence
techniques. McAllister et ul. (1983) found that mRNA
for egg laying hormone (ELH) and related peptides in the
opisthobranch Aplyxia californica is located in neurose-
cretory bag cells located adjacent to the abdominal (in-
testinovisceral) ganglion, and also in a small number of
central ganglia neurons. In veligers. cells containing this
mRNA are "distributed throughout the entire length of
the inner surface of the body wall, with one particularly
dense cluster of cells expressing ELH-related mRNA along
the body cavity close to the head." From this description,
it cannot be determined if presumptive ELH-containing
neurons, destined for central ganglia, are indeed located
within typical ectodermal proliferative zones for these
ganglia. The bag cell neurons, unlike neurons forming
initial primordia of central ganglia in developing gastro-
pods, do not begin to ingress from the ectoderm until
after metamorphosis and their definitive position is out-
side the abdominal ganglion proper. Therefore, the ob-
servations of McAllister el al. (1983) do not contradict
the notion that primordia of CNS ganglia arise initially
from stereotypical ectodermal locations on the veliger
body.
Pre-trochal cephalopedal ganglia
Derivation of cerebral ganglia from pre-trochal (in-
travelar) ectoderm, specifically the cephalic plates, was
first described in early cell lineage studies of gastropods
(see Conklin, 1897) and has been confirmed by many
subsequent analyses (reviewed by Raven, 1958; Moor,
1983; Verdonk and van den Biggelaar, 1983). As suggested
by Thompson (1958), imagination of the cephalic plates
during cerebral gangliogenesis may fulfill a need for in-
creased ectodermal surface area during rapid mitoses of
cephalic plate cells.
Tardy (1970. 1974) has stated that the invaginated ce-
phalic plates in the nudibranch Aeolidiella alderi are the
source of both cerebral and pleural neurons. According
to his interpretation, the ganglia formed from the two
cephalic plates are fused cerebropleural ganglia, each hav-
ing a dorsal lobe corresponding to the pleural component
and a ventral lobe corresponding to the cerebral com-
ponent. In young Melibe larvae, each developing cerebral
ganglion also has a small dorsal protuberance, but these
develop into the optic ganglia. The pleural ganglia of Me-
libc develop from a pair of post-trochal ectodermal plac-
odes, as described in the following paper.
Presumptive neurons within the cephalic plate ecto-
derm ofMelihe ingress singly or in small clusters through-
out the larval phase; they do not separate en masse from
the ectoderm during later development as suggested by
Tardy (1970: 'telencephalization') for the nudibranch
Acolidiella alderi.
I 'isceral loop and osphradial neurons
The fiber tract of the visceral loop forms a complete
circuit beneath the esophagus from the time Melibe larvae
hatch from the egg mass. This fiber tract is probably es-
tablished by axons from cerebral neurons, because many
differentiated cerebral neurons are present at the hatching
stage and the cerebral neuropiles are continuous with the
visceral loop fiber tract. With the exception of several os-
phradial neurons and one neuron associated with the right
pallial placode, differentiated neurons are not associated
with the visceral loop at the hatching stage. A visceral
loop arising directly from the cerebral ganglia is found in
Caudofoveata. Solenogastres. Monoplacophora. and
Polyplacophora, and may be the ancestral state for the
molluscan nervous system (Salvini-Plawen, 1985).
(iuiixliu identifications
In Melibe larvae, all primordial ganglia associated with
the visceral loop originate from visceropallial ectoderm
and show some degree of torsional displacement. In pro-
sobranchs, these characteristics apply to the intestinal and
visceral ganglia, with the osphradium and osphradial gan-
glion also arising from visceropallial ectoderm. The pleural
ganglia of prosobranchs ingress from post-trochal ce-
phalopedal ectoderm and do not show torsional displace-
ment (Smith. 1935; Crofts, 1937; Guyomarc'H-Cousin,
1974; Honegger, 1974; Demian and Yousif, 1975). Moritz
( 1939) is alone in describing a slight asymmetrical posi-
tioning for pleural ganglia in Crcpitlu/a ac/i/nca. I must
NUDIBRANCH NEUROGENESIS I
359
Figure 22. Longitudinal section through a larva after completing mantle retraction showing supraintestinal
ganglion (SPG) projecting dorsally from nght cerebral ganglion (CG). N = nephrocyst; RPP = right pallial
placode; ST = statocyst. Scale, 5 ^m.
Figure 23. Longitudinal section through same larva showing concentration of subintestinal neurons
(SBG) projecting ventrally from left cerebral ganglion (CG). N = ciliated tube of nephrocyst; ST = statocyst.
Scale. 5 Mm.
conclude that pleural neurons are not associated with the
visceral loop of Melibe and, therefore, are not part of the
adult 'cerebropleural' ganglia.
A pentaganglionate visceral loop (not including the
pleural ganglia), consisting of paired parietal ganglia in
addition to paired intestinal and unpaired visceral gan-
glia, has been called a major synapomorphy of opis-
thobranchsand pulmonates(Haszprunar, 1985b, 1988),
even though five distinct visceral loop ganglia are rarely
found among adults of this group. It is assumed that
ganglionic fusions have masked the pentaganglionate
condition in most Euthyneura. In support of this. Tardy
(1970, 1974) and Regondaud (1961, 1964) claimed to
resolve five visceral loop ganglia in a transient devel-
opmental stage of a nudibranch opisthobranch and ba-
sommatophoran pulmonate, respectively. However, the
visceral loop of developing euthyneurans is short and,
in Melibe at least, passes through a prolonged stage in
which cells are distributed along the entire left limb. I
argue that individual CNS ganglia cannot be defined
solely by the criterion of an apparent local concentration
of neuronal cells, but each must have a distinct site of
neuronal ingression from the ectoderm. Only in this way
can homologous ganglia be recognized in developmental
stages of different gastropods. Using this criterion, which
usually requires ultrastructural examination, I found
only four ganglionic ingression sites along the visceral
loop of Melibe larvae: subintestinal neurons arise from
the left pallial placode; visceral neurons arise from the
visceral placode: and two neuronal ingression sites were
resolved for the right pallial placode, one of which must
be the source of supraintestinal neurons (see below).
The neurons that differentiate from the right mantle
fold ectoderm, close to its shell-secreting periphery, can
be identified as osphradial by their position and because
they are linked to the right limb of the visceral loop by a
nerve tract. There are many descriptions of an osphradium
or osphradial ganglion in larvae of prosobranchs (Thiriot-
Quievreux, 1974: Demian and Yousif, 1975; others re-
viewed by Fretter and Graham, 1962) and opisthobranchs
that retain a mantle cavity through metamorphosis
(Smith, 1967; Kriegstein, 1977a, b). However, osphradial
neurons have not been identified previously in nudibranch
larvae, although Kempf el al. (1987) found a neuron close
to the edge of the right mantle fold in the nudibranch.
Tritonia diomedea. that labelled with a monoclonal an-
tibody to small cardioactive peptide B.
From the time of hatching, the osphradial neurons of
Melibe are associated with a mucous cell, which suggests
that the structure is homologous to the osphradial sensory
360
L. R. PAGE
NUDIBRANCH NEUROGENES1S I
361
epithelium of other gastropods. If this is the case, then is
there an osphradial ganglion? Present data are insufficient
to answer this question, but two possibilities can be sug-
gested. First, if the osphradial ganglion of gastropods is
generated from neurons that ingress from the osphradial
sensory epithelium, as proposed by Demian and Yousif
(1985), then this process does not occur during the on-
togeny of Melibe and there is no osphradial ganglion.
Consequently, the two ingression sites within the right
pallial placode can be interpreted as the source of right
parietal and supraintestinal neurons, respectively. Alter-
natively, these two ingression sites may generate neurons
homologous to those of the supraintestinal and osphradial
ganglia in other gastropods, in which case there is no ev-
idence of a right parietal ganglion. The latter possibility
may seem unlikely, and yet patelloid limpets (archeogas-
tropods) have a streak of sensory epithelium along the
distal mantle skirt that is connected by a nerve to a much
more proximally located osphradial ganglion with over-
lying osphradium (Haszprunar, 1985a). Comparative de-
velopmental studies on opisthobranchs having both os-
phradium and osphradial ganglion could help resolve this
uncertainty.
It might be argued that the more anterior of the two
neuronal ingression sites within the right pallial placode
is better identified as the source of right pleural neurons.
However, this interpretation would require that neurons
of the left pleural ganglion arise from the left pallial plac-
ode, which is obviously in a post-torsional location on
the right, ventro-lateral side of the larva. No other ecto-
dermal placode is associated with the left limb of the vis-
ceral loop in any larval stage that I examined. Further-
more, neurons that accumulate behind the left cerebral
ganglion project ventrally, reflecting their post-torsional
heritage. Therefore, the alternative interpretation is un-
acceptable because pleural ganglia are not affected by tor-
sion in gastropods (see Fretter and Graham, 1 962; Bullock,
1965). It follows that pleural ganglia must be derived from
ectoderm of the cephalopedal mass and their morpho-
genesis is described in the companion paper.
A.svmmelry oj the visceral loop
The larval digestive tract of Melibe is torted by less
than 90°, so partial torsion of visceropallial neurogenic
placodes is expected. However, the visceropallial placodes
exhibit differing amounts of torsion. The left pallial plac-
ode, which generates subintestinal neurons, is actually lo-
cated on the right side in a position showing marked tor-
sional displacement; it is far removed from the left neph-
rocyst. By contrast, torsional displacement of the right
pallial placode (and the osphradial neurons) is minimal,
particularly in young veligers. This placode is located im-
mediately adjacent to the right nephrocyst. The visceral
placode, like the digestive tract, exhibits approximately
60° of torsion. Consequently, all sites of pallial neurogenic
ectoderm are located toward the right side of the larva,
and the visceral loop is asymmetrical but never actually
streptoneurous. A similar pattern of non-uniform torsion
of the visceral loop was described by Crofts (1937, p. 250)
for larvae of the archaeogastropod Haliotis tuberculata.
following the initial (90°) phase of torsional twisting.
Regondaud (1961, 1964) found that the visceral loop of
Lynwaea stagnalis (Pulmonata) is only partially torted
and its asymmetrical trajectory is comparable to that of
Melibe larvae.
Cephalization and detorsion
Cephalization in gastropods is the anterior concentra-
tion of ganglia, particularly those of the visceral loop. Eu-
thyneury is the uncrossing of the twisted visceral loop and
displacement of ganglia to their pre-torsional sides; it is
essentially a reversal of torsion as it affects the nervous
system. Cephalization and euthyneury are believed to be
prominent trends among opisthobranch and pulmonate
gastropods (reviewed by Bullock, 1965; Schmekel, 1985).
Naef (1911) appears to have originated the idea that
euthyneury is due to shortening of the visceral loop. He
suggested that shortening concentrates the ganglia in the
head, thereby pulling them out of the body area subject
to torsion. This notion was more recently reiterated by
Tardy (1970). Naefs theory is false because observations
Figures 24-27. Series of frontal sections through larva at onset of mantle retraction showing merger of
right and left limbs of visceral loop and left pallial and visceral placodes. Orientation arrows: A = anterior;
P = posterior. L = left; R = right.
Figure 24. Right limb of visceral loop (R) extending toward left from mantle gland (MG). A = anus.
Scale. 5 urn.
Figure 25. Merger of right and left limbs of visceral loop (R and L. respectively) adjacent to anus (A)
and left pallial placode (LPP). Scale, 5 pm.
Figure 26. Left limb of visceral loop (VL) associated with left pallial placode (LPP). Section grazes
through periphery of visceral placode (VP) adjacent to terminal intestine (I). E = esophagus; MA = mantle;
ST = statocyst. Scale. 10 ^m.
Figure 27. Left pallial placode (LPP) with mitotic cell (arrow) and visceral placode (VP). I = intestine;
MA = mantle; VL = visceral loop. Scale, 5 >jm.
362
L. R. PAGE
29
Figure 28. Frontal section through larva at mantle fold hypertrophy stage showing merger of right and
left limbs of visceral loop (R and L, respectively) adjacent to intestine (I). MG = mantle gland; VG = visceral
ganglion; VP = visceral placode. Scale, 5 //m.
Figure 29. Slightly deeper section showing visceral nerve (arrows) extending around distal intestine (1)
from visceral ganglion (VG) towards visceral placode. Scale, 5 tim.
Figure 30. Visceral nerve (arrow) within base of visceral placode (VP). I = intestine. Scale. I ^m.
on Melibc and Lyninaea stagnalis (Regondaud, 1961.
1964) show that visceral loop ganglia differentiate from
torted visceropallial ectoderm, as they do in prosobranchs.
Nevertheless, the concept of visceral loop shortening as
the cause of euthyneury is popular (see Fretter and Gra-
ham, 1962; Bulloch, 1965). It might be envisioned that
visceral loop ganglia become concentrated against the ce-
rebral ganglia of their respective pre-torsional sides as
threaded beads pushed together during shortening of a
convoluted string. However, as pointed out previously by
Regondaud ( 1 96 1 ) for Lymnaea. and also seen in Melibc
and Aplysia califomica (Kriegstein, 1977a, b), all visceral
loop ganglia differentiate within close proximity along a
short visceral loop. This loop simply remains short in
cephalized species, while the rest of the body elongates
after metamorphosis. Instructions to elongate are given
instead to peripheral nerves. This is more than a semantic
point if cephalization is to provide a mechanism for eu-
NUDIBRANCH NEUROGENESIS 1
363
thyneury. because visceral loop ganglia of Mel the and
Lymmieii are close together and torted at the outset of
gangliogenesis.
£S
ISPG
33
I propose that euthyneury in Melibe can be explained
by two factors. First, half of the hypothesized detorsion
is fictional because ganglia differentiate from visceropallial
ectoderm that never shows more than partial torsion.
Second, existing evidence suggests that subintestinal neu-
rons, which arise from visceropallial ectoderm showing
marked torsional displacement (unlike supraintestinal
neurons), move toward the left cerebral ganglion by active
migration along the visceral loop fiber tract. Other evi-
dence that neurons migrate along existing connectives
during gastropod neurogenesis comes from the study of
McAllister el al. (1983), who found evidence that bag cells
migrate along the visceral loop in juveniles of Aplysia
californica. Active, leftward migration of subintestinal
neurons contributes to both cephalization and detorsion
without involving a length change of the visceral loop.
Similarly, visceral neurons appear to migrate along the
nerve linking visceral placode and visceral loop during
later larval development. However, it is unlikely that ac-
tive migration of neurons is responsible for the final dis-
placement of the visceral ganglion to the left side during
metamorphosis, because the movement is rapid and in-
volves all visceral neurons simultaneously. Instead, meta-
morphic movement of the visceral ganglion may be ac-
complished by differential lengthening of various con-
nectives and commissures linking the CNS ganglia.
Lengthening is restricted to the pedal and parapedal com-
missures and the segment of visceral loop extending be-
tween the supraintestinal and visceral ganglia (elongation
of the latter is still minimal compared to that which occurs
after metamorphosis in Aplysia californica). Therefore,
when the larval body expands rapidly after shell loss, the
visceral ganglion is pulled to the left because its tether to
the left side of the CNS remains short, whereas that con-
necting it to the right side lengthens. Displacement of the
visceral ganglion to the left side 'overshoots' detorsion.
Conclusions
Results of this study suggest that the paired 'cerebro-
pleural' ganglia of Melihe leonina are, in reality, a fusion
Figures 31-33. Histological cross sections through metamorphic
stages showing movement of visceral ganglion (arrow) from right to left
sides of post-larva. Orientation arrows: L = left; R = right. BG = buccal
ganglia: E = esophagus; CSBG = left cerebral + subintestinal ganglia;
CSPG = right cerebral + supraintestinal ganglia; PG = pedal ganglia;
ST = statocyst. Scales. 25 ^m.
Figure 31. Immediately after shell loss.
Figure 32. Five hours after shell loss.
Figure 33. Ten hours after shell loss. Visceral ganglion has fused
with left cerebral + subintestinal ganglionic complex. Arrowhead indicates
visceral loop fiber tract.
364
L. R PAGE
of cerebral, subintestinal. and visceral ganglia on the left
side, and cerebral, supraintestinal, and possibly parietal
and osphradial ganglia on the right side.
The basic configuration of the post-metamorphic CNS
and the pattern of peripheral nerves among many non-
dorid nudibranchs are generally similar to that of Mel i be
leonina. Among dendronotaceans, Dorsett (1978) ex-
tended this to include similarities at the level of individual
neurons, despite differences in adult body form. Therefore,
my conclusions about the identity and fate of visceral
loop ganglia may apply to many other non-dorid nudi-
branchs. However, small differences in adult neuroanat-
omy may signify large morphogenetic differences. Even
among members of the genus Melibe, Gosliner ( 1 987) has
documented differences in relative size and surface texture
of CNS ganglia. Development of visceral loop ganglia in
aeolids such as Phestilla sibogae, in which the dorsum
and cerata are derived from epipodial ectoderm rather
than pallial ectoderm (Bonar and Hadfield. 1974; Bonar,
1976), may be quite different from that in Melihe.
Dorid nudibranchs form a cohesive group with features
that distinguish them from other nudibranchs (Minichev,
1970; Schmekel, 1985). including anatomical details of
the CNS and pattern of peripheral nerves. Ganglionic re-
gions in the CNS of dorids requires separate neurodevel-
opmental analysis.
Acknowledgments
I am grateful for the encouragement and support pro-
vided by Dr. G. O. Mackie, who financed this study with
a grant from the Natural Science and Engineering Re-
search Council of Canada.
Literature Cited
Bickell, L. R., and F. S. Chia. 1979. Organogenesis and histogenesis
in the planktotrophic veliger of Dondclla steinhergae (Opisthobran-
chia: Nudibranchia). Mm Biol. 52: 291-313.
Bickell, I,. R., and S. C. Kempf. 1983. Larval and metamorphic mor-
phogenesis in the nudibranch Mcliht- lamina (Mollusca: Opisthob-
ranchia). Bint. Bull. 165: 119-138.
Bonar, I). B. 1976. Molluscan metamorphosis: a study in tissue trans-
formation. Am. Zool 16: 573-591.
Bonar, D. B., and M. G. Hadfield. 1974. Metamorphosis of the marine
gastropod Phestilla sihogac Bergh (Nudibranchia: Aeolidacea). I. Light
and electron microscopic analysis of larval and metamorphic stages.
/ £.v/>. Mar Bint. Ecol. 16: 227-255.
Bullock, T. H. 1965. The Mollusca. Pp. 1273-1515 in Structure ami
Function in ihc Neirous Systems ot Invertebrates. v.2, T. H. Bullock
and G. A. Horndge. eds. Freeman Press. San Francisco.
Conklin, E. G. 1897. The embryology of Crepidula. J Morphol 13:
1-226.
Creek, G. A. 1951. The reproductive system and embryology of
the snail Pnmalias elegans (Muller). Proc. Zoo/. Soc. Land. 121:
599-640.
Crofts, D. R. 1937. The development of //<;/;om tuhercitlalti. with spe-
cial reference to the Organogenesis during torsion. Phil. Trans. R
Soc. Land B228: 219-268.
Cumin. R. 1972. Normentafel zur Organogenese von Limnaea stagnalis
(Gastropoda. Pulmonata) mit besonderer Berlicksichtigung der Mit-
teldarmdruse. Rev. Snisse /MO/ 79: 709-774.
D'Asaro, C. N. 1969. The comparative embryogenesis and early or-
ganogenesis of Bursa cornigata and Distorsia clathrata. Malacologia
9: 349-389.
Demian, E. S., and F. Yousif. 1975. Embryonic development and or-
ganogenesis in the snail Marisa cornuarietis (Mesogastropoda: Am-
pullanidae). V. Development of the nervous system. Malacologia
15: 29-42.
Dorsett, D. A. 1978. Neurons controlling foot and mantle movements
in Arminti cali/ornica Mar Behav. Physiol 5: 307-324.
Dorsett, D. A. 1986. Brains to cells: the neuroanatomy of selected gas-
tropod species. Pp. 101-187 in The Mollusca. K. M. Wilbur, ed. V.9,
Newobiology and Behavior, pan 2. A. O. D. Willows, ed. Academic
Press. New York.
Fretter, V., and A. Graham. 1962. British Prosohranch Molluscs: Their
Functional Anatomy and Ecology. Ray Society. London.
Gosliner, T. M. 1981. Origins and relationships of primitive members
of the Opisthobranchia (Mollusca: Gastropoda). Biol. J. Linn. Soc.
16: 197-225.
Gosliner, T. M. 1987. A review of the genus Melibe (Opisthobranchia:
Dendronotacea) with descriptions of two new species. I'eliger 29:
400-414.
Gosliner, T. M. 1991. Morphological parallelism in opisthobranch
gastropods. Malacologia 32: 313-327.
Gosliner, T. M., and M. Ghiselin. 1984. Parallel evolution in opis-
thobranch gastropods and its implications for phylogenetic meth-
odology. Sysl /(><)/. 33: 255-274.
Guyomarc'H-Cousin, C. 1974. Etude descriptive de 1'organogenese du
systeme nerveux chez Litlonna saxatilis (Olivi) Gasteropode Pro-
sobranche. Ann Emhryol. Morphog. 4: 349-364.
Has/prunar. G. 1985a. The fine morphology of the osphradial sense
organs of the Mollusca. 1 . Gastropoda, Prosobranchia. Phil Trans.
R Soc Lond B307: 457-496.
Haszprunar, G. 1985b. The Heterobranchia— a new concept of the
phylogeny of the higher Gastropoda. Zeitsc/i. Zoo/. Svst. Evol. 23:
15-37.
Haszprunar, G. 1988. On the origin and evolution of major gastropod
groups, with special reference to the Streptoneura. J. Moll. Stud 54:
367-441.
Hoffmann. H. 1936. Opisthobranchia. Pp. 641-864 in Klassen initi
Ordungen des Tierrcichs. H. G. Bronns, ed. Akad. Verlagsges., Leipzig.
Honegger, T. 1974. Die Embryogenese von Ampiillarim (Gastropoda,
Prosobranchia). Zoo/. Ja/ir. Abl Anal Ontog. Ticrc93: 1-76.
kandel, E. R. 1979. Behavioral Biology of Aplysia. Freeman. San
Francisco.
Kempf, S. C., B. Masinovsky, and A. O. D. Willows. 1987. A simple
neuronal system characterized by a monoclonal antibody to SCP
neuropeptides in embryos and larvae of Tritonia diomedea (Gastro-
poda, Nudibranchia). / Neurobiol 18: 217-236.
Kriegslein, A. R. 1977a. Development of the nervous system of Aplysia
calilonuca. Proc. Nail Acad. Set. USA 74: 375-378.
Kriegstein, A. R. 1977b. Stages in the post-hatching development of
Aplysia calilonuca J E\p. Zool 199: 275-288.
McAllister, L. B., R. H. Scheller, E. R. Kandel, and R. Axel. 1983. In
sun hybridization to study the origin and fate of identified neurons.
Science 222: 800-808.
Minichev, Y. S. 1970. On the origin and system of nudibranchiate
molluscs (Gastropoda, Opisthobranchia). Monilore Zool. llal. (N.S.)
4: 169-182.
NUD1BRANCH NEUROGENESIS I
365
Minichev, V. S., and V. I. Starobogatov. 1978. On the systematic ar-
rangement of euthyneuran snails. Malacol. Rev 11: 67-68.
Moor, B. 1983. Organogenesis. Pp. 123-177 in The Mollinca, K. M.
Wilbur, ed.. v.3. Development. N. H. Verdonk. J. A. M. van den
Biggelaar. and A. S. Tompa. eds. Academic Press. New York.
MiirilA C. K. 1939. Organogenesis in the gasteropod Crepulula adiinca
Sowerby. L'nn Call! Berkeley Publ. Zool 43: 217-248.
Naef, A. 1911. Studien zur generellen Morphologic der Mollusken. 1.
Ueber Torsion und AsymetriederGasteropoden. /•.'/>,'/> l-'onsch. Zool.
3: 73-164.
Page, I.. R. 1992. New interpretation of a nudibranch central nervous
system based on ultrastructural analysis of neurodevelopment in
Melibe leonina II. Pedal, pleural, and labial ganglia. Biol. Bull. 182:
366-38 1 .
Raven, C. P. 1958. Morphogenesis: the Analysis ofMolluscan Devel-
opment Pergamon Press. Oxford.
Regondaud, J. 1961. Formation du systeme nerveux et torsion chez
Lymnaea stagnalis L. (Mollusque Gasteropode). C. R Hebd. L 'Acad.
Sci. 252: 1203-1205.
Regondaud, J. 1964. Origine embryonnaire de la cavite pulmonaire
de Lymnaea stagnalis L. Considerations particulieres sur la mor-
phogenese de la commissure viscerale. Bull Bioi Fr Belgium1 98:
433-471.
Richardson. K. C., L. Jarrett, and E. H. Finke. 1960. Embedding in
epoxy resins for ultrathin sectioning in electron microscopy. Stain
Techno/. 35: 313-323.
Russell, I.. 1929. The comparative morphology of the elysioid and
aelidioid types of the molluscan nervous system, and its bearing on
the relationships of the ascoglossan nudibranchs. Proc. Zool. Soc.
Loud. 14: 197-233.
Salvini-Plawen, L. v. 1985. Early evolution and the primitive groups.
Pp. 59-150 in The Mollusca. K. M. Wilbur, ed., v. 10. Evolution.
E. R. Trueman and M. R. Clarke, eds. Academic Press, New York.
Schacher, S., E. R. Kandel, and R. YYoolley. 1979a. Development of
neurons in the abdominal ganglion of Ap/ysia californiea. 1. Axo-
somatic synaptic contacts. Dev Biol. 71: 163-175.
Schacher, S., E. R. Kandel, and R. Woolley. 1979b. Development of
neurons in the abdominal ganglion of Ap/ysia californiea. II. Non-
neural support cells. Dev. Biol. 71: 176-190.
Schmekel, L. 1985. Aspects of evolution within the Mollusca. Pp. 221-
267 in The Mollusca, K. M. Wilbur, ed., v.10. Evolution. E. R. True-
man and M. R. Clarke, eds. Academic Press, New York.
Smith, F. G. \V. 1935. The development of Patella vulgala Phil. Trans.
R Sue I.ond. S225: 95-125.
Smith, S. T. 1967. The development of Retusa obtusu (Montagua)
(Gastropoda, Opisthobranchia). Can. J. Zool. 45: 737-764.
Tardy, J. 1970. Contribution a 1'etude des metamorphoses chez les
nudibranches. Ann Sci. Nat Zool. Pans 12: 299-370.
Tardy, J. 1974. Morphogenese du systeme nerveux chez les mollusques
nudibranches. Haliotis 4: 61-75.
Thiriot-Quievreu\, C. 1974. Anatomic interne de veligeres plancto-
mques de Prosobranches Mesogastropodes au stade proche de la me-
tamorphose. Thai Jugoslav. 10: 379-399.
Thompson, T. E. 1958. The natural history, embryology, larval biology,
and post-larval development of Ada/ana proximo (Alder and Han-
cock) (Gastropoda, Opisthobranchia). Phil Trans R Soc. Loud.
5242: 1-58.
Verdonk, N. H., and J. A. M. van den Biggelaar. 1983. Early devel-
opment and the formation of the germ layers. Pp. 91-122 in The
Mollusca. K. M. Wilbur, ed., v.3, Development. N. H. Verdonk,
J. A. M. van den Biggelaar, and A. S. Tompa, eds. Academic Press,
New York.
Willows, A. O. D. ed. 1985-1986. The Mollusca. K. M. Wilbur, ed..
v.8 and 9, Neurobiology and Behavior, parts 1 and 2. Academic Press,
New York.
Reference: Biol. Bull 182: 366-381 (June, 1992)
New Interpretation of a Nudibranch Central Nervous
System Based on Ultrastructural Analysis
of Neurodevelopment in Melibe leonina.
II. Pedal, Pleural, and Labial Ganglia
LOUISE R. PAGE
Department of Biology. University of Victoria. Victoria. British Columbia. Camula \'$}V 2Y2
Abstract. Electron microscopical analysis of semi-serial
sections through larval stages of the dendronotid nudi-
branch Melibe leonina (Gould, 1852) revealed paired
placodes of neurogenic ectoderm at the base of the foot.
The location of these laterocephalic placodes corresponds
to descriptions of the ectodermal site generating pleural
neurons in prosobranchs. In Melihe, there are two sites
of neuronal ingression within each laterocephalic placode.
Neurons ingressing from one of these sites join the cerebral
ganglia, and their initial axons extend into the cerebro-
buccal connectives or run distally along the esophagus. I
identify these neurons as homologues of labial ganglia
neurons in archeogastropods. However, neurons derived
from the second ingression site within each laterocephalic
placode join the pedal ganglia. Pedal ganglia are present
in hatching veligers and are linked to the cerebral ganglia
by cerebropedal connectives associated with the statocyst
nerves. A second connective between each cerebral and
pedal ganglia appears at the onset of neuronal ingression
from the laterocephalic placodes. Peripheral axons
branching from this second pair of connectives are asso-
ciated with laterocephalic neurons that ingress to the pedal
ganglia. I argue that these are pleural neurons, meaning
that the pleural ganglia in Melibe are uncoupled from the
visceral loop.
Introduction
The nervous systems of opisthobranch gastropods have
proven highly amenable to neurophysiological investi-
gations (reviewed in Kandel, 1979; Willows, 1985-1986),
and have been used to support or criticize phylogenetic
theories for this taxonomically difficult subclass (Guiart,
Received 6 November 1991; accepted 5 March 1992.
1901; Russell, 1929; Boettger. 1954; Gosliner. 1981;
Haszprunar, 1985, 1988; Schmekel, 1985). As a result,
the neuroanatomy and neurophysiology of this group is
the subject of a large body of literature, with the former
studies extending back to the last century. It might be
expected that the basic structure of opisthobranch central
nervous systems (CNS) would be thoroughly understood.
In fact, homologous ganglionic regions within the often
highly consolidated nervous systems of opisthobranchs
are essentially best guesses based on comparisons of adult
neuroanatomy in primitive and derived species.
In some opisthobranchs. distinct pleural ganglia are
linked by connectives to ipsilateral cerebral and pedal
ganglia, and are also the first pair of ganglia along an
elongate visceral loop bearing additional ganglia. This ar-
rangement conforms to the basic design for the gastropod
nervous system (reviewed by Bullock, 1965; Dorsett,
1986). However, separate pleural ganglia are not distin-
guishable in most extant opisthobranchs. In nudibranchs,
it is always assumed that the pleurals have fused with the
cerebral ganglia to form a pair of cerebropleural ganglia
(see Guiart, 1901; Hoffmann, 1936; Boettger, 1954:
Schmekel, 1985). This interpretation seems entirely logical
because the visceral loop enters the posterior lobes of the
•cerebropleural' ganglia, and each 'cerebropleural' gan-
glion is often linked to the ipsilateral pedal ganglion by
two connectives (Fig. 1 ). Presumably, the more posterior
of these two connectives is the pleuropedal. and the an-
terior connective is the cerebropedal. However, in my
previous study of gangliogenesis in the nudibranch Melihe
leonina. I found no evidence of pleural ganglia associated
with either the visceral loop or cerebral ganglia. All ganglia
of the visceral loop in Melihe arise from ectoderm ot the
visceropallium and show torsional displacement, whereas
366
NUDIBRANCH NEUROGENESIS II
367
Figure 1. Light micrograph of the central nervous system of Melibe
Icoinna at 48 h after metamorphic shell loss (lateral view). Traditionally,
the two ganglia have been called cerebropleural (CPLG) and pedal ganglia
(PG). Note the two connectives (large arrowheads) with statocyst (ST)
associated with posterior connective. EY = eye. Scale 25 ^m.
the pleural ganglia of gastropods are not affected by tor-
sion.
If the pleural ganglia are not part of the visceral loop
in Melibe, then where are they? Ultrastructural study of
semi-serial thin sections through larval stages of this nu-
dibranch. suggests that the pleural ganglia are fused with
the pedal ganglia. This interpretation differs from that of
an earlier developmental study by Bickell (now Page) and
Kempf (1983), based on histological sections. Evidence
of labial ganglia in larvae of Melibe was a second unex-
pected result. These are distinct ganglia in some archaeo-
gastropods and a pyramidellid, but not in other adult gas-
tropods (Fretter and Graham, 1949, 1962). The labial
ganglia in larval Melibe fuse with the ventral side of the
cerebral ganglia prior to metamorphosis.
All members of the genus Melibe are characterized by
an oral hood, an expansion of the circumoral cephalic
epidermis that is used to capture prey (Gosliner. 1987).
This structure appears to be homologous to the smaller
oral veil of other dendronotacean nudibranchs. Nev-
ertheless, the large size of the oral hood may have modified
gangliogenesis in Melibe, relative to that in other nudi-
branchs, particularly ganglia arising from cephalopedal
ectoderm. Therefore, the generality of this new model for
nudibranch CNS structure must be tested by further stud-
ies on other species.
Materials and Methods
Methods for rearing and anaesthetizing larvae of Melibe
leonina were described by Bickell and Kempf ( 1983): the
fixation method was that of Bickell and Chia (1979). The
technique for semi-serial thin sectioning and the larval
stages examined were described in the first paper of this
duet (Page, 1992).
Results
As justified in the previous paper (Page, 1992), gangli-
onic primordia were identified by the locations of their
neurogenic ectoderm, as compared to those described for
the ganglia of prosobranchs. Trajectories of associated
connectives, commissures, and peripheral axon tracts were
also very useful for identifying ganglia derived from post-
trochal cephalopedal ectoderm. The sketch in Figure 2
shows approximate positions of neurogenic ectoderm for
pedal, pleural, and labial ganglia in veligers that have
completed mantle retraction (approximately midway
through the larval phase).
Pedal ganglia and cerebropedal connectives
In hatching veligers, an axon tract emerges from the
ventral aspect of each cerebral ganglion, extends past the
ipsilateral statocyst, and associates with a small cluster of
subepidermal pedal cells. The axon tracts are the left and
right cerebropedal connectives and the bilaterally sym-
metrical cell clusters within the foot are anlagen of the
pedal ganglia (Fig. 3). The statocyst nerves, which each
leave their respective statocyst in company with a blind,
ciliated static duct, combine with the ipsilateral cerebro-
pedal connective as the two join the cerebral ganglion
(Fig. 4). Therefore, the cerebropedal connectives can be
LCP
SH
Figure 2. Right lateral view of Melibe leonina larva shortly after
mantle retraction showing developing CNS (scant stippling) and ap-
proximate locations of neurogenic ectoderm (dense stippling). The lat-
erocephalic placodes (LCP) include zones of ingression for both labial
and pleural neurons (LN and PLN, respectively). CG = cerebral gang-
lion; CP = cephalic plate; EV = eye; F = foot; MA = mantle fold;
PG = pedal ganglion; S = stomach; SH = shell; SPG = supraintestinal
ganglion; ST = statocyst; V = velar lobe; VL = visceral loop.
368
L R PAGE
identified in all larval and post-larval stages by their as-
sociation with the statocysts and statocyst nerves. In young
larvae, the cerebropedal connective is the only axon tract
extending between each cerebral and pedal ganglion. In
post-metamorphic animals, the cerebropedal connective
becomes the more posterior of two connectives extending
between each of these ganglia (Fig. 1 ).
Within each pedal ganglion of hatching veligers, the
fiber tract of the cerebropedal connective breaks up into
small bundles (Fig. 5) that extend independently to the
overlying pedal ectoderm.
During subsequent development, the pedal ganglia en-
large greatly by addition of cells ingressing from paired
ectodermal placodes (pedal placodes) extending down
each side of the ventral surface of the foot. Sites of cellular
ingression appear to be restricted to where the peripheral
axon tracts run into the pedal ectoderm (Fig. 6). These
peripheral tracts become the anterior, medial, and pos-
terior pedal nerves of post-metamorphic animals.
The left and right pedal ganglia become connected by
a pedal commissure by mantle retraction stage, and a par-
apedal commissure is distinguishable in histological sec-
tions of metamorphic animals.
Pleura! ganglia
In 6-day-old larvae, a pair of thickened ectodermal
placodes has appeared where the two sides of the foot
merge with the head, just beneath the origin of the velar
lobes and lateral to the mouth and statocysts (Figs. 7, 8,
9). The placode on the right side is larger than that on
the left, but the positions of the two are bilaterally sym-
metrical (Fig. 7). These laterocephalic placodes are im-
mediately proximal to the pedal placodes that give rise to
pedal ganglion neurons, and the right laterocephalic plac-
ode is separated from the right pallial placode by the
nephrocyst pores. Therefore, the laterocephalic placodes
are components of post-trochal cephalopedal ectoderm.
On each side, the visceral loop and cerebropedal con-
nective travel past the ipsilateral laterocephalic placode,
shortly after these fiber tracts emerge from their respective
cerebral ganglion (Fig. 9). However, neither tract associates
in any way with the adjacent laterocephalic placode in 6-
day-old larve. Placodal ectodermal cells do not begin to
ingress until shortly before mantle retraction.
Sections through larvae at 6 days after hatching suggest
that the base of each laterocephalic placode is constricted
by underlying muscle fibers (Fig. 9). The adjacent mem-
branes of both the muscle and placodal cells are thrown
into folds (Fig. 10). and the two are connected by nu-
merous adherens junctions. Furthermore, the manner in
which the ectodermal cells are contorted suggests that their
basal ends are being pulled toward the pedal placode lo-
cated more distally along the foot (Fig. 8).
Each laterocephalic placode eventually delivers neurons
to the developing CNS from two separate sites of cellular
ingression. Neurons arising from the more ventro-medial
of these sites fuse with the ipsilateral cerebral ganglion;
these are homologues of labial ganglia (see following sec-
tion of Results). Neurons ingressing from the second, more
lateral site fuse with the ipsilateral pedal ganglion and
represent the pleural ganglion on each side. Onset of cel-
lular ingression from the laterocephalic placodes is cor-
related with the appearance of the second connective ex-
tending between each cerebral and pedal ganglion. I iden-
tify these as the cerebropleural connectives; they lie
anterior to the previously formed cerebropedal connec-
tives. The positions of the labial and pleural ganglia in
Mc'libe larvae, relative to other components of the CNS,
is shown in Figure 1 1.
For an unknown reason, ingression of cells from the
two laterocephalic placodes does not occur symmetrically
on both sides of the larva. Labial neurons begin ingression
on the left side before ingression starts on the right side.
Conversely, the timetable for ingression of pleural neurons
on the right side may be slightly ahead of that on the left
side, although this appearance may actually result from
fewer pleural neurons ingressing from the left compared
to the right placode (the left laterocephalic placode is
smaller than the right). Nevertheless, in larvae examined
after full mantle retraction, concurrent ingression of both
labial and pleural neurons is evident on both sides.
The series of micrographs in Figures 12 to 15 show an
early stage in the development of the right pleural ganglion
and reveal its relationship to the cerebropleural connective
Figure 3. Low magnification eleclronmicrograph of slightly oblique cross section through newly hatched
larva ofAIelibe Icunina showing right pedal ganglion (PG: enlarged in Fig. 5) located distal to the statocysts
(ST) within the foot (F), and the upper extremity of the right pallial placode (RPP). E = esophagus; MA =
shell secreting cells of mantle; N = nephrocysts. Scale 10 ^m.
Figure 4. Cross section through base of right cerebral ganglion (CG) in newly hatched larva showing
close association between statocyst nerve (SN) and cerebropedal connective (CPC) where they merge with
the cerebral neuropile (NP). SD = static duct. Scale 3 /jm.
Figure 5. Pedal ganglion (PG) of newly hatched larva beneath ectoderm of foot (F). Arrows indicate
three small axon bundles. Scale 3 ^m.
Figure 6. Portion of pedal ganglion (PG) and overlying pedal placode (PP; label is on mitotic cell) at
mantle fold hypertrophy stage. Ingressing pedal neurons (arrowheads) are associated with peripheral axon
tract (arrow). Scale 3 ^m.
NUDIBRANCH NEUROGENESIS I!
369
370
L. R. PAGE
and pedal ganglion. Figure 12 passes through the latero-
cephalic placode adjacent to the statocyst. An enlarged
area of this section (Fig. 13) reveals that the right cere-
bropedal connective, associated with the statocyst nerve,
has emerged from the cerebral ganglion, whereas the fiber
tract of the cerebropleural connective is still associated
with the antero-ventral extremity of the cerebral ganglion.
In Figure 14, the cerebropleural connective is leaving the
cerebral ganglion and the fiber tract is bifurcating. In Fig-
ure 15, a bulge of neurons is evident on the anterior border
of the pedal ganglion; this is the anlage of the right pleural
ganglion. The cerebropedal connective is entering the
pedal ganglion and the main body of the cerebropleural
fiber tract is deflecting medially. Axons forming the other
branch of the previous bifurcation of the cerebropleural
connective are emerging from the pleural ganglion at two
sites. These peripheral axon tracts extend to the adjacent
laterocephalic placode and are associated with ingressing
pleural neurons (Fig. 15). Later in larval development, at
least one of the two peripheral axon tracts extending to
each laterocephalic placode is joined by axons that appear
to arise from the cerebropedal connective.
The cerebropleural connective lengthens after its initial
formation and the boundary between pleural and pedal
ganglia is indistinguishable in late stage larvae.
Labial neurons have not begun to ingress from the right
laterocephalic placode at onset of mantle retraction, al-
though this process is evident on the left side as described
in the following section.
Labial ganglia
The series of micrographs shown in Figures 16 to 19
were taken from the left side of the same larva shown in
Figures 12 to 15. They illustrate the early formation of
the left labial ganglion from an ingression site within the
laterocephalic placode that is distinct from the ingression
site for pleural neurons.
In Figure 16, the cerebropedal connective has left the
cerebral ganglion but the cerebropleural connective is still
within the antero-ventral extremity of this ganglion. Im-
mediately after the cerebropleural connective has left the
cerebral ganglion, the ganglion acquires a prominent lat-
eral extension of neurons (Fig. 17) that is continuous with
the laterocephalic placode (Fig. 19). These neurons are
clearly fusing with the cerebral ganglion, and evidence
from a slightly older developmental stage (see below) in-
dicates they are homologues of the labial ganglia found
in archeogastropods.
Figures 17 and 18 show a second site of neuronal
ingression from the laterocephalic placode, located slightly
beneath and lateral to that for labial neurons. These are
ingressing pleural neurons associated with a small tract
of axons that branched from the cerebropleural connec-
tive. The pleural neurons are extending toward the pedal
ganglion, but unlike the right pedal ganglion of this larva,
the left pedal ganglion has not yet acquired a prominent
bulge of pleural neurons.
The micrographs in Figures 20 to 25 show the trajec-
tories of initial axons elaborated by labial ganglion neurons
on the left side. At the time of fixation, this larva had
completed mantle retraction and mantle fold cells had
begun proliferation prior to forming the definitive dorsal
epidermis. Figure 20 passes through an area comparable
to that of the younger stage shown in Figure 1 6. In addition
to the cerebropedal and cerebropleural connectives, the
cerebrobuccal connective is prominent and extends to a
thickening of the ventral esophageal wall that is neurogenic
ectoderm for the left buccal ganglion. A subsequent sec-
tion through the base of the cerebrobuccal connective (Fig.
2 1 ) shows a group of axons that originated from the fiber
tract of the cerebrobuccal connective, plus two other axon
tracts that both originated from labial neurons (Figs. 22,
23). One of the labial axon tracts merges with the tract
extending from the cerebrobuccal connective (Fig. 21),
whereas the other labial tract extends directly into the
cerebral ganglion (not shown). The combined labial and
cerebrobuccal axons form a peripheral nerve that extends
distally along the wall of the esophagus (Fig. 22), forming
synapses on esophageal cells just inside the mouth (labial
nerve; Figs. 24. 25).
Animals sectioned immediately after metamorphic shell
loss show a prominent plug of neuronal cell bodies within
the ventro-lateral aspect of each cerebral ganglion, anterior
to the connectives. This corresponds to the site where
labial neurons have joined the cerebral ganglia. The labial
Figure 7. Low magnification electron micrograph of a cross section through a 6-day-old larva. Boxed
areas contain left and right laterocephalic placodes (LCP), which are enlarged in Figures 8 and 9, respectively.
Note size difference between two placodes. Orientation arrows: L = left; R = right; D = dorsal; V = ventral.
E = esophagus; M = mouth; N = nephrocysts; ST = statocysts. Scale 10 nm.
Figure 8. Left LCP. Note how bases of placodal cells are flexed in direction of foot (arrow). M = muscle
fiber. Scale 3 ijm.
Figure 9. Right LCP. Note adjacent cerebropedal connective (CPC) and visceral loop (VL) emerging
from cerebral ganglion Muscle fibers(M) underly placodal cells. Asterisk marks position of nephrocyst duct
in slightly deeper section. MF = right mantle fold. Scale 3 pm.
Figure 10. Base of right LCP. Arrowheads indicate processes from underlying muscle fiber (M) extending
to convoluted basal lamina of placodal ectoderm. Scale 1 urn.
NUDIBRANCH NEUROGENESIS II
"" '' • '•.'*?
371
,.v , .• •.>.
. •'
372
L. R. PAGE
CPC
Figure 11. Reconstruction of cephalopedal ganglia and supramtes-
tinal ganglion in larva of Mel i he Iconina. right, ventro-lateral view. Gan-
glia and connectives labelled on right side only. For clarity, rhinophoral
ganglia are omitted and connectives over the statocysts are elongated.
Broken lines show initial axon tracts of labial neurons (all tracts do not
necessarily arise from single neurons). BG = huccal ganglion; CBC
= cerehrobuccal connective: CC = cerebral commissure: CG = cerebral
ganglion; CPC = cerebropedal connective: CPLC = cerebropleural con-
nective; E = esophagus; EY = eye; LG = labial ganglion; PG = pedal
ganglion; PLG = pleural ganglion; ST = statocyst: VL = visceral loop.
nerve becomes the nerve labelled Cl in adult Mclibc by
Hurst (1968). This nerve innervates the oral tube and lips.
Discussion
Identifications of ganglia
Pedal ittuiglui. The pedal ganglia in Mclibc Iconina,
like those in other gastropods, arise from proliferative ec-
toderm along ventro-lateral zones of the larval foot. In-
gressing pedal neurons are associated with axons that ex-
tend from the intraganglionic fiber tract of each cerebro-
pedal connective, to the overlying neurogenic ectoderm
of the pedal placode. These axon tracts become peripheral
nerves after metamorphosis. A similar association between
ingressing neurons and peripheral axon tracts was ob-
served for pleural and visceral ganglia.
In Aplysia californica, the number of neurons within
all central ganglia continue to increase after metamor-
phosis (Cash and Carew, 1989), and results of McAllister
ct ul. (1983) and Hickmott and Carew ( 199 1 ) suggest that
added neurons come from the body wall. Jacob (1984)
and McAllister cl ul. ( 1983) proposed that ingressing neu-
rons in larvae and juveniles of this species migrate along
connective tissue strands or muscle fibers to reach their
definitive locations within the CNS. Alternatively, obser-
vations on Melibc larvae suggest that ingressing neurons
may be guided to developing ganglia by migrating along
peripheral axon tracts. After metamorphosis, these tracts
continue to connect the CNS with the often distant body
wall, and are therefore ideally suited to guide later in-
gressing neurons to appropriate central ganglia.
Pleural ganglia. My identification of pleural ganglia in
Mclibe larvae, and their developmental fate, is probably
the most controversial part of this analysis and therefore
requires detailed justification.
Histological studies of neurogenesis in prosobranchs
and pulmonates have shown that the ectodermal placode
giving rise to pleural neurons is post-trochal (Smith. 1935;
Crofts, 1937; Regondaud, 1961, 1964; D'Asaro, 1969;
Cumin. 1972; Honegger. 1974; Guyomarc'H Cousin,
1974; Demian and Yousif. 1975; Raven. 1975). Nev-
ertheless, Tardy (1970, 1974), studying the nudibranch
Aeolidiella uklcri by means of histological sections,
claimed that both cerebral and pleural neurons are derived
from pre-trochal cephalic plate ectoderm, so that cerebral
and pleural ganglia are fused from the outset. Smith ( 1967)
suggested the same for the cephalaspid Ret lisa obtiisa,
and Jacob (1984) also claimed a common site of origin
for cerebral and pleural ganglia in Aplysia calijbniicu.
If previous studies have correctly identified a post-tro-
chal ectodermal ingression site for pleural neurons in pro-
sobranchs and pulmonates. then I reject the notion of a
common origin for cerebral and pleural neurons in opis-
thobranchs. Conklin's (1897) monumental studv of cell
Figures 12 to 15. Series of frontal sections through right laterocephalic placode (LCP) showing cerebro-
pleural connective and pleural ganglion.
Figure 12. Right LCP immediately beneath velar lobe (V) and opposite the statocyst (ST). Boxed area
enlarged in Figure 13. CG = cerebral ganglion; CP = cephalic plate; PG = periphery of pedal ganglion.
Scale 5 ftm.
Figure 13. Detail from Figure 1 2 showing cerebropedal connective (CPC) with associated statocyst nerve
>N): cerebropleural connective (CPLC) is still within cerebral ganglion (CG). Scale 2 ^m.
Figure 14. Same area in slightly deeper section showing cerebropleural connective (CPLC) emerging
from cerebral ganglion (CG). Arrow indicates axons branching from CPLC. Scale 2 urn.
Figure 15. Subsequent section showing anlage of pleural ganglion (PLG) perched atop pedal ganglion
(PG). Note liber tracts of cerebropedal and cerehropleural connectives (CPC and CPLC, respectively). Axons
that branched from CPLC in Figure 14 are emerging from pleural ganglion at two points (arrows). Note
pleural neurons (PL) ingressing from laterocephalic placode (LCP). CG = cerebral ganglion; ST = statocyst.
Scale 2 fim.
NUDIBRANCH NEUROGENESIS II
eV. '-'^l
\ »«•>**..• 'ilyfti nA.^*
373
ST
15
374
L R. PAGE
lineage in the prosobranch gastropod, Crepidula, showed
that the cephalic plates are derived from the first quartet
of embryonic micromeres, whereas post-trochal ectoderm
is derived from subsequent micromere quartets. This has
been confirmed by many other studies (reviewed by Ra-
ven, 1958; Verdonk and van den Biggelaar. 1983). in-
cluding Casteel's (1904) cell lineage study on the nudi-
branch Fiona marina. Therefore, none of the neurons
ingressing from the cephalic plate in opisthobranchs can
be homologous to pleural neurons ingressing from post-
trochal ectoderm in other gastropods.
The ectodermal proliferation placode for pleural neu-
rons must be located within the cephalopedal mass, rather
than the visceropallial mass, because the pleural ganglia
of gastropods do not show torsional displacement. The
exact location of the paired ectodermal placodes for the
pleural ganglia in a variety of prosobranchs has been de-
scribed as: the sides of the head, base of the foot, opposite
the statocysts or within the pleural groove. Even in Lit-
torina saxatilis, a caenogastropod with an epiathroid adult
nervous system (pleural ganglia close to cerebrals), the
pleural ganglia arise from the base of the foot (Guyo-
marc'H-Cousin, 1974). In Melibe, the location described
by these phrases corresponds to the post-trochal, latero-
cephalic placodes.
Jacob (1984), who used 3H-thymidine autoradiography
to study neurogenesis in Aplysia californica, suggests that
pleural neurons arise from the velar lobes. However, the
velar proliferation placode shown in her Figure 4a is
clearly the large ciliated cells of the prototroch, with the
subvelar ridge (metatroch) immediately below. What Ja-
cobs (1984) interprets as velar cells migrating to the ce-
rebral, and ultimately the pleural ganglia (see her Fig. 5a).
may be the area where the prototroch arches over the
mouth, closely approaching the cerebral ganglia (see Fig.
4 in the companion paper. Page, 1992). However, Jacob's
figure 5b also shows 'H-thymidine labelling in a protrud-
ing placode of ectodermal cells, lateral to the statocyst,
that corresponds in appearance and location to a latero-
cephalic placode in Melibe larvae.
Appropriate location within post-trochal cephalopedal
ectoderm is only one of several clues that help identify
the ectodermal ingression site for pleural neurons in Me-
libe larvae. Another is the correlation between the onset
of cellular ingression from this site and the appearance of
the second pair of connectives extending in front of the
statocysts from the cerebral ganglia. These connectives
are clearly not the earlier formed cerebropedal connectives
associated with the statocyst nerve. Furthermore, the two
peripheral axon tracts extending to each laterocephalic
placode branch from this second pair of connectives,
which are presumably the cerebropleural connectives.
These two axon tracts may correspond to the anterolateral
and dorsal 'pedal' nerves described in adult Melibe rosea
by De Vries (1963). Based on peripheral projection pat-
terns, De Vries ( 1963) believed that the anterolateral nerve
in Melibe rosea represents part of the ancestral anterior
pleural nerve.
The size difference between left and right laterocephalic
placodes correlates with a similar size difference between
left and right dorsal 'pedal' nerves in adults of Melibe and
other non-dorid nudibranchs. This difference, in turn,
may relate to unilateral innervation of the penis by the
right member of this nerve pair. Similarly, Regondaud
(1961) noted that the right-side ectodermal placode for
generating pleural neurons in Lymnaea xtagnalis is larger
than that of the left side. The pleural ganglia in Lymnaea
are associated with the visceral loop and do not fuse with
other ganglia during development.
I cannot readily explain the slight timing differences
for developmental events involving the left and right lat-
erocephalic placodes. Possibly, the temporal asymmetries
are somehow related to the size difference between the
two placodes and the unilateral innervation of the penis
from the right side. Whatever the reason, similar devel-
opmental asynchronies involving bilaterally homologous
structures are not unusual among mollusks, even for
structures not affected by torsion. Examples include the
cephalic tentacles of some prosobranchs, the statocysts of
Bnccinnm. and the ctenidia of the bivalve, Ostrea (re-
viewed by Moor, 1983).
The slight temporal differences for developmental
events involving the left and right laterocephalic placodes
are dwarfed by the much greater difference between the
onset of neuronal ingression from the pedal placodes and
that from the laterocephalic placodes. Presumptive neu-
Figures 16 to 19. Series of frontal sections through left side of larva at onset of mantle retraction showing
labial ganglion developing from left laterocephalic placode.
Figure 16. Cerebropedal connective (CPC) associated with statocyst nerve (SN). and cerebropleural
connective (CPLC; enlarged in inset) still within anteroventral extremity of cerebral ganglion (CG). CP
= cephalic plate; ST = statocyst. Scale 5 jim; inset 0.5 /<m.
Figure 17. Subsequent section showing left LCP and labial ganglion (LG| projecting laterally from
cerebral ganglion (CGI. Arrow indicates cerebropleural connective, which is enlarged in inset. Ingressing
pleural neurons (PL) are enlarged in Figure 18. PG = pedal ganglion; ST = statocyst; V = velar lobe. Scale
5 /jm; inset 0.5 j/m.
Figure 18. Ingressing pleural neurons (PL) associated with axons (arrowheads). Scale 5 fim.
Figure 19. Arrow indicates mitotic labial ganglion neuron ingressing from LCP. CG = cerebral ganglion;
PG = pedal ganglion; ST = statocyst; V = velar lobe. Scale 5 ^m.
NUDIBRANCH NEUROGENESIS II
375
376
L. R. PAGE
rons begin to ingress from the pedal placodes even before
hatching, where, s laterocephalic neurons do not begin to
leave the eel Dderm until the latter part of the shell-se-
creting phase. This marked difference further attests to
the separate identity of neurons arising from laterocephalic
and pedal placodes.
The notion that the pleural ganglia are primarily pallial
in nature has become rather entrenched in the malacolog-
ical literature. Although these ganglia certainly contribute
to the innervation of the mantle in adult gastropods, they
are not derived developmentally from visceropallial ec-
toderm, and Crofts ( 1937) observed that the pleural nerves
("external pallial nerves") extending into the mantle folds
in Haliotis are not established until the long-lasting second
phase of torsion is complete. If this were not the case, the
trajectory of these nerves would be skewed by torsion.
If my arguments are accepted for the identity of the
visceral loop fiber tract and associated ganglia (Page,
1992), and for the pleural ganglia in Mclibe larvae, then
one must also accept a conclusion that is unprecedented
in previous developmental or neuroanatomical studies of
gastropods: the cell bodies of the pleural neurons in Mclihe
have become uncoupled from the visceral loop. This con-
clusion is actually not so startling, because the visceral
loop appears to be established by axons extending from
the cerebral ganglia, with other ganglia applied later. A
visceral loop (lateral cords), arising directly from the ce-
rebral ganglia, is thought to be the ancestral condition of
the gastropod nervous system and is exhibited by mono-
placophorans, caudofoveates, solenogastres, and polypla-
cophorans (see Bullock. 1965; Salvini-Plawen, 1985;
Wingstrad, 1985); none of these groups have pleural gan-
glia. In Alelibe, and probably all gastropods, it is inap-
propriate to think of the visceral loop as an initiative of
the pleural ganglia. Furthermore, uncoupling of pleural
ganglia from the visceral loop is no less bizarre than the
complex restructuring of the visceral loop and ganglia that
occurs during development of ampullarid prosobranchs
(Honegger, 1974; Demian and Yousif, 1975).
Circumstantial support for the reality of pleuropedal
ganglia among other dendronotid nudibranchs comes
from neuroanatomical and neurophysiological data on
Tritonia diomedea and T. hombergi (Willows et ai, 1973;
Dorsett, 1974), and on Armina californica (Dorsett, 1978).
The dorsum and branchial tufts (cerata) of tritoniids are
derived from larval mantle fold ectoderm (Thompson,
1962; Kempf and Willows, 1977). Mantle fold ectoderm
also gives rise to the intestinal and visceral ganglia, which
fuse to the cerebral ganglia, so it is not surprising that
some neurons effecting branchial tuft withdrawal are lo-
cated within the lobes projecting posteriorly from the ce-
rebral ganglia (traditionally called the pleural lobes).
However, two other effector neurons for branchial tufts
are located in each 'pedal' ganglion. Extrapolating from
the fact that axons of pleural neurons project into pallial
tissues in shelled gastropods, it is possible that these neu-
rons have a pleural ganglion ancestry. It is interesting to
note that one of these neuronal pairs (L and RPdl), par-
ticularly that of the right side, is occasionally found within
the ipsilateral cerebral ganglionic mass; that is, in the ex-
pected position if pleural ganglia were not uncoupled from
the visceral loop. During the development of Alelibe, I
found that both the visceral loop and the pedal ganglia
are positioned equidistant from the laterocephalic plac-
odes in 6-day-old larvae. A relatively minor developmental
miscue might easily deflect individual pleural neurons to-
wards the visceral loop, rather than the pedal ganglion.
Morphological evidence suggests that contracting muscle
fibers underlying the laterocephalic placodes may pull
placodal cells toward the pedal ganglion, and therefore
away from the visceral loop fiber tract.
As in tritoniids, Dorsett (1978) found neurons in the
pedal ganglia of Aniiinu culitimiicii that control move-
ments of the 'mantle' periphery (this species lacks cerata).
Figures 20 to 25. Series of frontal sections through left side of larva at mantle fold hypertrophy stage
showing anlage of labial ganglion (LG) and trajectory of labial axons. Sections proceed distally along esophagus
toward mouth.
Figure 20. Anteroventral portion of left cerebral ganglion (CG) with developing cerebrobuccal connective
(large arrowheads) extending to buccal placode (BP) in ventral wall of esophagus (E). CPC = cerebropedal
connective; CPLC = cerebropleural connective; PG = periphery of pedal ganglion; SN = statocyst nerve;
ST = statocyst. Scale 5 ijm.
Figure 21. Cerebral ganglion end of cerebrobuccal connective showing two bundles of labial axons
irrows), one of which is extending toward axons of cerebrobuccal connective (CBC). Scale 1 ^im.
Figure 22. Arrow indicates axons arising from labial ganglion neurons (LG; enlarged in Fig. 23). Note
nerve (LN) adjacent to wall of esophagus (E). Scale 3 urn.
Figure 23. Detail from Figure 22 showing an axon (arrowheads) arising from a labial ganglion neuron
(LG). Scale 0.5 Mm.
Figure 24. Section passing just inside mouth showing labial ganglion (LG) continuous with laterocephalic
placode (LCP). Arrowheads indicate ingressing pleural neurons. Boxed area, enlarged in Figure 25. includes
labial nerve. E, esophagus; PPG = pleuropedal ganglion. Scale 5 Mm.
Figure 25. Detail from Figure 24 showing labial nerve with probable synapse (arrowhead) onto esophageal
cell (E). Scale 0.5 Mm.
NUD1BRANCH NEUROGENESIS II
377
378
I.. R. PAGE
Some of these, or other pedal neurons, also control move-
ments of the foot or oral veil. These efferents can be stim-
ulated by tactile .stimuli to the foot or mantle margins.
Walters el il (1983) described a conspicuous cluster
of mechanosensory neurons within the undisputed pleural
ganglia ofAplysia californica. Collectively, these primary
mechanosensory neurons are sensitive to tactile stimuli
applied along the entire length of the foot, including the
'parapodia' (epipodial folds). Therefore, the receptive field
for pleural sensory neurons in Aplysia corresponds ap-
proximately to the location of larval neurogenic ectoderm
for presumed pleural neurons in Melibe.
Labial ganglia. Unlike pleural neurons, ingressing labial
neurons fuse with the ipsilateral cerebral ganglion in Me-
libe. Separate labial ganglia, which are linked beneath the
buccal mass by a labial commissure and which give rise
to labial nerves innervating the oral lips and buccal mus-
culature, are found in some archaeogastropods and a pyr-
amidellid (Fretter and Graham. 1949, 1962). They are
thought to be absent or fused to the cerebral ganglia in
other gastropods.
In archaeogastropods such as Haliotis, the cerebrobuc-
cal connectives are routed through the labial ganglia before
arriving at the buccal ganglia (see Fretter and Graham,
1962). This design can be recognized in the relationship
between cerebrobuccal and labial nerve tracts in larval
Melibe, and is the major justification for labelling these
ganglia as homologues of archaeogastropod labial ganglia.
Audesirk (1979) and Audesirk and Audesirk (1980) in-
vestigated two adjacent clusters of mechanoreceptive
neurons within each cerebral ganglion of Tritonia diome-
dea. The two clusters have distinct axonal projections and
neurophysiological characteristics, but both are located
where the labial ganglia are fused to the cerebral ganglia
in Melibe (ventro-lateral side of each cerebral ganglion,
anterior to the connectives to the pedal ganglia). One of
the clusters sends axons out each ventral cerebral nerve
(labelled CN4 in Tritonia but clearly equivalent to CN1
of Melibe) and also into the adjacent cerebrobuccal con-
nective; a trajectory that is consistent with my observations
on initial labial axons in Melibe larvae. These mecha-
noreceptors respond to pressure or stretch of the oral tube
or jaw closer muscles (Audesirk, 1979). The second cat-
egory of mechanoreceptive neurons sends axons out ce-
rebral nerves extending to the oral veil and a pedal nerve
extending to the anterior foot (Audesirk and Audesirk.
1980). Although further data are needed, 1 consider it
very possible that i . • ebral nerves innervating the oral veil
in Tritonia and the oral hood in A fell be are also established
by labial neurons.
It might be argued that the labial ganglia are better
identified as the pleural ganglia in Melibe. They arise from
a pair of post-trochal ectodermal sites of appropriate lo-
cation for pleural neurons and they fuse with the cerebral
ganglia, which the traditional interpretation dictates for
the pleural ganglia of opisthobranchs. However, the tra-
jectory of axons arising from labial neurons is inappro-
priate for pleural neurons.
Critique of previous newodevelopmental models
All previous accounts of sequential neurodevelopment
in nudibranchs have been histological, yet this method
cannot reveal with certainty all neuronal ingression sites
from ectodermal proliferation placodes or trajectories of
early connectives and axon tracts. Uncertainty resulting
from limited resolution has resulted in three different
models (more including variations) for the pattern of
neurogenesis in nudibranchs. These earlier models, plus
the current model generated from ultrastructural study of
Melibe leonina larvae, are illustrated schematically in
Figure 26 and are discussed below.
The groundbreaking study of nudibranch organogenesis
was done by Thompson ( 1958) on the lecithotrophic ve-
liger of the dorid nudibranch, Adalaria proximo.. Thomp-
son identified cerebral, optic, pedal, pleural, and buccal
ganglia in Adalaria (Fig. 26A). The intestinal and visceral
ganglia escaped notice, possibly because they may fuse
precociously to the cerebrals in this lecithotrophic veliger.
Alternatively, Tardy (1970) has suggested that Thompson
misidentified visceral loop ganglia as the buccal ganglia.
According to Thompson ( 1958), the pleural ganglia of
Adalaria are large neuronal masses located within the base
of the foot, and they fuse with the cerebrals at metamor-
phosis. The 'anterolateral propodial ganglia' described by
Chia and Koss (1989) and Arkett el at. (1989) for the
dorid nudibranch Onchidoris bilamellata have a similar
size and location to the ganglia identified as the pleurals
in Adalaria. Further study is needed to determine if these
are indeed pleural ganglia, or alternatively, the labial gan-
glia.
Tardy (1970, 1974) proposed that both cerebral and
pleural neurons arise from the cephalic plate in Aeolidiella
alder:, thereby producing cerebropleural ganglia that are
fused from the beginning (Fig. 26B). As discussed earlier,
this contradicts observations made on many other species.
Tardy also recognized buccal ganglia in Aeolidiella alderi.
but his description of their early development is highly
anomolous. Many studies, including this one on Melibe,
have found that neurogenic ectoderm for buccal ganglia
is located within the ventral wall of the distal esophagus
(Smith, 1935; Creek, 1951; D'Asaro, 1969;Guyomarc'H-
Cousin, 1974; Honegger, 1974; Demian and Yousif,
1975), but Tardy (1974; p. 315 and Fig. 2E) shows the
location of buccal neurogenic ectoderm as lateral to the
mouth, proximal to the pedal placodes. and adjacent to
the statocysts. I suggest that these are the laterocephalic
placodes. Tardy also described a pair of ganglia that arise
from the ectoderm of the oral tentacles, but only after
metamorphosis.
NUDIBRANCH NEUROGENESIS II
379
Having rejected Tardy's (1970, 1974) interpretation
for the origin of pleural ganglia in nudibranchs, my
colleagues and I identified the ganglia directly behind
the cerebrals as the pleural ganglia in previous
histo. logical studies on the dorid Doridella steinbergae
(Bickell and Chia, 1979) and on Melibe leonina (Bickell
and Kempf, 1983). Kempf et al. (1987) gave the same
interpretation for the larval CNS of the dendronotid
Tritonia diomedea, except they also resolved a visceral
ganglion (Fig. 26C). This model is superficially attrac-
tive because it produces an adult CNS with ganglionic
regions that conform to the traditional interpretation
based on comparative neuroanatomical studies of adult
opisthobranchs (Fig. 27A). However, this model fails
to explain why the left 'pleural ganglion' in the larva
projects ventrally relative to the right, when pleural
ganglia should not be involved in torsional displace-
ments. 1 now believe that these ganglia are the intes-
tinals.
Figure 26D shows my current interpretation of gangli-
onic regions within the late larval nervous system of Me-
libe leonina. The subsequent pattern of ganglionic fusions
produces a CNS with ganglionic regions that differ from
the traditional interpretation. The differences are illus-
trated in Figure 27.
-PLG
OTG
SBG-
Figure 26. Four interpretations of the developing nudibranch CNS.
Relative lengths of connectives and sizes of ganglia are not accurate;
optic and rhinophoral ganglia not shown. (A) Thompson ( 1958) for .(</-
alaria proxima. (B) Tardy ( 1970, 1974) for Acoiuhclln Men (C) Bickell
and Chia ( 1 979) for Doridella steinhergaeand Bickell and Kempf ( 1983)
for \ feli he leonina: visceral ganglion added by Kempf el al (1987) for
Tritonia diomedea. (D) present study on Melihe leunimi. Abbreviations:
BG = buccal ganglia; CG = cerebral ganglion; LG = labial ganglion;
LPAG = left parietal ganglion; OS = osphradial neurons; OTG = oral
tentacle ganglion; PG = pedal ganglion; PLG = pleural ganglion: RPAG
= right parietal ganglion; SBG = subintestinal ganglion: SPG = suprain-
testinal ganglion; VG = visceral ganglion.
VL
LG
VL
Figure 27. Interpretations of ganglionic regions in adult CNS ofAfe-
lihe leonina (optic and rhinophoral ganglia not shown). (A) Traditional
Model. Arrowhead no. 1 indicates cerebropedal connective; arrowhead
no. 1 indicates pleuropedal connective. (B) Revised Model. Arrowhead
no. 1 indicates cerebropleural connective; arrowhead no. 2 indicates ce-
rebropedal connective. BG = buccal ganglion; CC = cerebral commissure;
CG = cerebral ganglion: EY = eye; LG = labial ganglion (ventralK
located); PG = pedal ganglion; PLG = pleural ganglion; SBG = subin-
testinal ganglion; SPG = supraintestinal ganglion: VG = visceral ganglion;
VL = visceral loop.
Thoughts on phytogeny
Questions about phyletic relationships among opis-
thobranchs and about opisthobranch origins continue to
be debated. Uncertainties stem from the large degree of
morphological variation within the group and many in-
stances of apparent parallelism (Ghiselin, 1965: Gosliner.
1981, 1991; Gosliner and Ghiselin, 1984). Nervous sys-
tems have a reputation for conservatism and are therefore
valued as phyletic characters. It might be hoped that opis-
thobranch nervous systems could help unmask relation-
ships otherwise hidden by superficial morphological dif-
ferences or convergences. To date, proposed phylogenies
have considered only adult neuroanatomical characters,
with emphasis given to the length and detorsion of the
visceral loop, and extent of ganglionic fusions. Using the
latter criterion. Russell (1929) argued convincingly that
nudibranchs are unlikely ancestors for sacoglossans. Nev-
ertheless, Ghiselin (1965). among others, believes that
380
L. R PAGE
most of these characters show polyphyletic trends, and
the extensively fused nervous systems of many adult opis-
thobranchs offer few other unambiguous clues to assist in
phylogenetic reconstructions. However, ultrastructural
study of neurodevelopment has revealed much more
about an opisthobranch nervous system than is apparent
from adult CNS structure. Indeed, the new developmental
data about homologous ganglionic regions in Mel/he raise
serious questions about the validity of adult-based neu-
roanatomical interpretations given for other opistho-
branchs.
In a revision of gastropod systematics. Salvini-Plawen
and Haszprunar (1987) and Haszprunar (1988) argued
that a hypoathroid or dystenoid nervous system (fused or
semi-fused pleural and pedal ganglia) is a diagnostic char-
acter for the archaeogastropod grade. Based partly on the
assumption that opisthobranch pleural ganglia are fused
to the cerebrals (epiathroid nervous system: typical of
caenogastropods), when not distinct in adults, Haszprunar
( 1985. 1988) proposed a primitive caenogastropod ances-
tor for the Opisthobranchia. Gosliner (1981) came to the
same conclusion based on a suite of many criteria. It is
therefore surprising to find a hypoathroid nervous system
in a nudibranch. although the pleural ganglia of Aplysia
califomica are also very close to the pedal ganglia (well
illustrated by fig. 4 of Cash and Carew, 1989). These ob-
servations contradict Haszprunar's proposed criterion for
the archaeogastropod grade and undermine the notion of
an epiathroid nervous system for all opisthobranchs. The
possibility that the condition in Mclihc is merely an ex-
ceptional, secondarily derived state, possibly related to
the unusual oral hood of this genus, must be determined.
I have identified pleural neurons in Mclihc, partly on
the assumption that the two connectives between "cerebral"
and 'pedal' ganglia are cerebropedal and cerebro-pleural-
pedal connectives. This dictum for gastropod CNS or-
ganization is supported by a large body of comparative
anatomical data, but its universality is certainly not
proven. Again, ultrastructural studies of development for
a variety of gastropod groups are needed.
To date, labial ganglia have been identified in adults
of some archaeogastropods (reviewed by Fretter and Gra-
ham. 1962). a pyramidellid (Fretter and Graham. 1949;
pyramidellids have strong opisthobranch affinities), and
in larvae of the nudibranch Mclihc lamina. However, if
the subccrebral commissure is evidence of labial ganglia
fused to cerebral ganglia, as suggested by Bullock ( 1965),
then labial ganglia are present also in other opisthobranchs
and in pulmonates. Neither labial ganglia nor a subcere-
bral commissure have been found among caenogastro-
pods, nor have they been reported in developmental stages
of this group.
In conclusion, the nervous system of Mclihc shares
major plesiomorphous characters (pleuropedal ganglia,
labial ganglia) with that of archaeogastropods. Although
the uncoupling of the pleural ganglia from the visceral
loop is an apomorphy not suspected previously for any
other gastropod, the Melibe nervous system is not readily
derived from the epiathroid nervous system typical of ex-
tant caenogastropods. The validity of these speculations
must await further comparative studies.
Acknowledgments
It is a pleasure to thank Dr. G. O. Mackie for encour-
agement and financial support, and Dr. D. H. Paul for
discussions of this topic. The study was funded by a grant
from the Natural Science and Engineering Research
Council of Canada to GOM.
Literature Cited
Arkctt, S. A., F. S. Chia, J. I. Goldberg, and R. Koss. 1989. Identified
settlement receptor cells in a nudibranch veliger respond to specific
cue. Biol. Bull 176: 155-160.
Audesirk, T. E. 1979. Oral mechanoreceptors in Triionia diomcdca
I . Electrophysiological properties and location of receptive fields. ./.
Comp Physml. A130: 71-78.
Audesirk, G., and I. K. Audesirk. 1980. Complex mechanoreceptors
in Trillium ilioincdea I. Responses to mechanical and chemical
stimuli../ Comp Phyuol AMI: 1(11-109.
Bickell, L. R., and F. S. Chia. 1979. Organogenesis and histogenesis
in the planktotrophic veliger of Doridella steinbergae (Opisthobran-
chia: Nudibranchia). Mar Biol 52:291-313.
Bickell. L. R., and S. C. Kempf. 1983. Larval and metamorphic mor-
phogenesis in the nudibranch Melibe Iconina (Mollusca: Opistho-
branchia). Biol Bull 165: 119-138.
Boettger, C. 195-4. Die Systematik der euthyneuren Schnecken. Zoo!.
An: Lf//>n.i,'(suppl.) 17: 253-279.
Bullock, T. 11. 1965. The Mollusca. Pp. 1273-1515 in Slriictwc ami
l-'i/iiLlion in the \cr\-ous Sy^cim i>l Inrenehrales. V. 2, T. H. Bullock
and G. A. Horridge, eds. Freeman Press, San Francisco.
Cash, D., and T. J. Carew. 1989. A quantitative analysis of the devel-
opment of the central nervous system in juvenile Aplysia califomica
J. i\citn )/>;«/. 20: 25-47.
Casteel, D. B. 19(14. The cell-lineage and early larval development of
Fiona manna, a nudibranch mollusc. Proc. Acad. Nal. Sei. Phila-
delphia 56: 325-405.
Chia, F. S., and R. Koss. 1989. The fine structure of a newly discovered
propodial ganglia of the veliger larva of the nudibranch Onchuloris
hilanu'lluici Cell Tissue Re-., 256: 17-26.
Omklin, E. G. 1897. The embryology of Crepidnhi ./ Morpliol 13:
1-226.
Creek, G. A. 1951. The reproductive system and embryology of the
snail Pomatias clcgans (Muller). Proc. /.ool Sue. Loud 121:
599_64().
Crofts, D. R. 1937. The development ofHaliolis inbcrcnlutti, with spe-
cial reference to the organogenesis during torsion. Phil Trans. R
Soc Lorn/ B228: 219-268.
Cumin. R. 1972. Normentalel zur Organogenese von Limnaea stagnalii
(Gastropoda. Pulmonata) mil besonderer Beriicksichtigung der Mit-
teldarmdruse. Rc\: Sinsse /ool 79: 709-774.
D'Asaro, C. N. 1969. The comparative embryogenesis and early or-
ganogenesis of Buna comixalu and Dislorsui elarlhrala Malaeolo^ia
9: 349-389.
Herman, E. S., and F. Yousif. 1975. Embryonic development and or-
ganogenesis in the snail Mansa eornuarielis (Mesogastropoda: Am-
NUDIBRANCH NEUROGENESIS II
381
pullariidae). V. Development of the nervous system. Malacologia
15: 29-42.
De Vries, J. B. 1963. Contribution to the morphology and histology
of the nudibranch Mclihe rosea Rang. Ann. i'mv. Stellenbosch 38
(ser. A): 105-153.
Dorset!, D. A. 197-4. Neuronal homologies and the control of branchial
tuft mo\ements in two species of Trilonia. J Exp. Biul. 61:
639-654.
Dorset!. D. A. 1978. Neurons controlling foot and mantle movements
in Armina califomica Mm Bc/niv Physiol 5:307-324.
Dorset!, D. A. 1986. Brains to cells: the neuroanatomy of selected gas-
tropod species. Pp. 101 -187 in The Mollusca. K. M. Wilbur, ed. V.9,
.\eurohiologr and Behavior, part 2. A. O. D. Willows, ed. Academic
Press. New York.
Frctler, \ ., and A. Graham. 1949. The structure and mode of life of
the Pyramidellidae. parasitic opisthobranchs. ./ Mar. Bid/. Assoc.
L'.K. 28:493-532.
Fretlcr, V., and A. Graham. 1962. British Prosobranclt Molluscs: Their
Fiinctiiiiuil Analomy and Ecology. Ray Society. London.
Ghiselin, M. 1965. Reproductive function and the phylogeny of opis-
thobranch gastropods. Malacologia 3: 327-378.
Gosliner, T. M. 1981. Onginsand relationships of primitive members
of the Opisthobranchia (Mollusca: Gastropoda). Biol. J. Linn. Soc.
16: 197-225.
Gosliner, T. M. 1987. A review of the genus Me/ibe (Opisthobranchia;
Dendronotacea) with descriptions of two new species. 1'eliger 29:
400-414.
Gosliner, T. M. 1991. Morphological parallelism in opisthobranch
gastropods. Malacologia 32: 313-327.
Gosliner, I'., and M. Ghiselin. 1984. Parallel evolution in opisthobranch
gastropods and its implications for phylogenetic methodology. Syst.
Zoo/. 33: 255-274.
Guiart, J. 1901 . Contribution a 1'etude des gasteropodes opisthobranchs
et en particulier de cephalaspides. Mem. Soc. Zoo/ Fr 14: 5-219.
Guyomarc'H-Cousin, C. 1974. Etude descriptive de I'organogenese du
systeme nerveux chez Lillorina saxatilis (Olivi) Gasteropode Pro-
sobranche. Ann. Embryol. Morphog. 4: 349-364.
Has/prunar, G. 1985. The Heterobranchia — a new concept of the
ph\logen\ of the higher Gastropoda. Z. Zoo/. Syst. Evol 23: 15-37.
llas/prunar. G. 1988. On the origin and evolution of major gastropod
groups with special reference to the Streptoneura. J. Moll. Stud 54:
367-441.
Ilickmoti, P. \V., and T. J. Carew. 1991. An autoradiographic analysis
of neurogenesis in juvenile Aplysia califomica. J. Neurobiol. 22: 3 1 3-
326.
Hoffmann. 11. 1936. Opisthobranchia. Pp. 641-864 in Klassen inul
Ordungen des Tierreichs. v.3, H. G. Bronns, ed. Acad. Verlagsges..
Leipzig.
Honegijer. T. 1974. Die Embryogenese von Ampul/arias (Gastropoda.
Prosobranchia). Zoo/. Jahrb Abl. Anal. Onlogenie Ticre93: 1-76.
Hurst, A. 1968. The feeding mechanism and behaviour of the opis-
thobranch Mchbc Iconina. Symp. Zool. Soc. Lond. 22: 151-166.
Jacob. M. H. 1984. Neurogenesis in Aplysia califomica resembles ner-
vous system formation in vertebrates. / Neurosei 4: 1225-1239.
Kandel, K. R. 1979. Behavioral Biology of Aplysia. Freeman, San
Francisco.
Kempf, S. C., B. Masinovsky, and A. O. D. Willows. 1987. A simple
neuronal system characterized by a monoclonal antibody to SCP
neuropeptides in embryos and larvae of Trilonia diomedea (Gastro-
poda. Nudibranchia). J. Ncurobiol. 18: 217-236.
Kempf, S. C., and A. O. D. Willows. 1977. Laboratory culture of the
nudibranch Trilonia diomcdea Bergh (Tritoniidae: Opisthobranchia)
and some aspects of its behavioral development. ./. Exp. Mar. Biol
Ecol. 30:261-276.
McAllister, L. B., R. II. Scheller. K. R. Kandel, and R. Axel. 1983. In
situ hybridization to study the origin and fate of identified neurons.
Science 222: 800-808.
Moor, B. 1983. Organogenesis. Pp. 123-177 in The Mollusca. K. M.
Wilbur, ed.. v.3. Development. N. H. Verdonk. J. A. M. van den
Biggelaar. and A. S. Tompa. eds. Academic Press. New York.
Page, L. R. 1992. New interpretation of a nudibranch central nervous
system based on ultrastructural analysis of neurodevelopment in
Melibe Iconina I. Cerebral and visceral loop ganglia. Biol Bull 182:
348-365.
Raven, C. P. 1958. Morphogenesis: The Analysis ot Molluscan Devel-
opment. Pergamon Press. Oxford.
Raven, C. P. 1975. Development. Pp. 367-400 in Pulmonalcs. Vol. I.
V. Fretter and J. Peake, eds. Academic Press, New York.
Regondaud, J. 1961. Formation due systeme nerveux et torsion chez
Lymihica slagnnlis L. (Mollusque Gasteropode). C.R. Hebd. Sceances
L'Acad. Sci. 252: 1203-1205.
Regondaud, J. 1964. Ongine embryonnaire de la cavhe pulmonaire
de Lymnaea stagnalis L. Considerations particulieres sur la mor-
phogenese de la commissure viscerale. Bull. Biol. Fr. Bclg 98: 433-
471.
Russell, L. 1929. The comparative morphology of the elysioid and
aeolidioid types of the molluscan nervous system, and its bearing on
the relationships of the ascoglossan nudibranchs. Proc. Zoo/. Sue.
Lond. 14: 197-233.
Salvini-Plawen, L. v. 1985. Early evolution and the primitive groups.
Pp. 59-150 in 77/i- Mollusca. K. M. Wilbur, ed., V.IO. Evolution.
E. R. Trueman and M. R. Clarke, eds. Academic Press, New York.
Salvini-Plawen, L. v., and G. Has/prunar. 1987. The Vetigastropoda
and the systematics of the streptoneurous gastropods (Mollusca). J.
Zool. Lond. A211: 747-770.
Schmekel, L. 1985. Aspects of evolution within the Mollusca. Pp. 221-
267 in 77;t' .Mollusca. K.. M. Wilbur, ed., v. 10. Evolution. E. R. True-
man and M. R. Clarke, eds. Academic Press, New York.
Smith, K. G. \V. 1935. The development of Patella vu/gala. Phil. Trans.
R Soc Lond 8225:95-125.
Smith, S. T. 1967. The development of Retitsa obtusa (Montagua)
(Gastropoda, Opisthobranchia). Can. J. Zoo/. 45: 737-764.
Tardy. J. 1970. Contribution a 1'etude des metamorphoses chez les
nudibranches. Ann. Sci. \al.. Zoo/.. Pans 12: 299-370.
Tardy, J. 1974. Morphogenese du systeme nerveux chez les mollusques
nudibranches. Halions 4: 61-75.
Thompson, T. E. 1958. The natural history, embryology, larval biology,
and post-larval development of . ida/aria pmxima (Alder and Han-
cock) (Gastropoda, Opisthobranchia). Phil Trans R Soc. Lond.
B242: 1-58.
Thompson. T. E. 1962. Studies on the ontogeny of Trilonia hombergi
Cuvier (Gastropoda Opisthobranchia). Phil Trans. R Soc. Lond.
B245: 172-218.
Verdonk, N. H.. and J. A. M. van den Biggelaar. 1983. Early devel-
opment and the formation of the germ layers. Pp. 91-122 in The
Mollitscd K M. Wilbur, ed., \.3. Development. N. H. Verdonk,
J. A. M. \an den Biggelaar. and A. S. Tompa. eds. Academic Press,
New York.
Walters, E. T., J. H. Byrne, T. J. Carew, and E. R. Kandel.
1983. Mechanoafferent neurons innervating tail of Aplysia. I. Re-
sponse properties and synaptic connections. J. Neurophysiol. 50:
1522-1542.
Willows, A. O. D., ed. 1985-1986. The Mo/lusca. K. M. Wilbur, ed.
v.8 and 9, \eiirobinlogy and Behavior, parts 1 and 2. Academic Press,
New York.
Willows, A. O. D., D. A. Dorset!, and G. Hoyle. 1973. The neuronal
basis of behavior in Trilonia 1 . Functional organization of the nervous
system. J \eurobiol 4: 207-237.
\\ingstrad. K. G. 1985. On the anatomy and relationships of recent
Monoplacophora. (iia/alhea Rep. 16: 7-94.
Reference: Biol Bull 182: 382-390. (June. 1992)
Control of Cilia in the Branchial Basket
of Ciona intestinalis (Ascidacea)
DWIGHT BERGLES* AND SIDNEY TAMM
Boston University Marine Program, Marine Biological Laboratory,
Woods Hole. Massachusetts 02543
Abstract. We investigated arrest and inactivation re-
sponses of stigmatal cilia in the branchial basket of the
ascidian, Ciona intestinalis. Using an improved prepa-
ration of living tissue for microscopic imaging of ciliary
responses, we found that Ca-ionophore A23187 in sea-
water + 50 mM Ca caused actively beating cilia to assume
the upright inactive posture, while A23187 in seawater
+ 100 mA/Ca caused transient (5-10 s) stigma-wide ar-
rests in which the cilia lie flat against the stigmatal walls.
Both responses are therefore Ca dependent, but the in-
active state has a lower threshold for Ca than does arrest.
Membrane permeant cAMP analogues induced >40% of
the quiescent cilia within a stigma to begin beating.
Saponin-extracted models of stigmatal cilia were de-
veloped to study the ionic and molecular control of ciliary
activity in Ciona. Extracted cilia were stimulated to beat
vigorously for >45 min by ATP-containing reactivation
solution (RS). Addition of 10~5 to 10"' M Ca to reacti-
vation solution caused the cilia to stand upright (inacti-
vate), but not to arrest. The calmodulin antagonists tri-
fluoperazine and calmidazolium (100 nM) restored active
beating when included in RS + 50-100 /iMCa, thereby
reversing Ca-dependent inactivation. Addition of bovine
brain calmodulin to RS + 1 00 nM Ca did not cause arrest
of reactivated cilia. RS + 100 ftM cAMP + 1 mM 3-
isobutyl-1-methyl-xanthine or the catalytic subunit of c-
AMP-dependent protein kinase increased both the pro-
Received 1 3 December 1991: accepted 27 March 1 992.
* Present address: Department of Molecular & Cellular Physiology.
Beckman Center. Stanford University School of Medicine. Stanford,
California 94305-5426.
Abhrcvmnony B-cAMP. N6-benzoyl-cAMP; M-cAMP. N6-mono-
butryl-cAMP; ES, extraction solution; WS. wash solution; RS. reacti-
vation solution; TFP, tnfluoperazine: IBMX, 3-isobutyl-l-methyl-xan-
thine; TAME, Na-p-tosyl-L-arginine methyl ester, PICA, c-AMP-depen-
dent protein kinase.
portion and vigor of reactivated beating. Addition of 100
M^/Ca to the RS + cAMP + IBMX solution caused reac-
tivated cilia to vibrate or twitch in an upright position,
suggesting that Ca and cAMP have antagonistic effects
on stigmatal cilia.
Introduction
The activity of cilia and flagella is regulated in many
organisms, reflecting the important role of these organelles
in locomotion, suspension feeding, gas exchange, mucous
and gamete transport, and sperm chemotaxis. Well-known
examples of ciliary and flagellar responses to stimuli in-
clude reorientation of beat direction in ciliate protozoa
(Eckert ct a/.. 1976; Machemer. 1986), arrest of lateral
cilia of lamellibranch gills (Murakami and Takahashi,
1975:Tsuchiya, 1977; Walter and Satir, 1978), inhibition
of velar cilia of molluscan larvae (Carter, 1926), activation
ofAfytilits abfrontal cilia (Stommel, 1984), Beroe macro-
cilia (Tamm. 1988). and sperm flagella (Brokaw. 1987),
changes in waveform of sea urchin sperm flagella (Brokaw
el u/-. 1974; Brokaw, 1979) and Chlamydonionas flagella
(Hyams and Borisy, 1978; Bessen ct a/.. 1980), and re-
versal of direction of wave propagation in trypanosome
flagella (Holwill and McGregor, 1976). Many of these ax-
onemal responses are known to be triggered by depolar-
ization-induced changes in intracellular Ca or alterations
in cyclic nucleotide levels (Machemer, 1986; Brokaw,
1987; Otter, 1989; Stephens and Stommel, 1989; Preston
and Saimi, 1990; Bonini ct ai. 1991).
Tunicates show periodic interruptions in the beating of
cilia that line openings (stigmata) of the branchial basket
and generate the feeding current (Fedele, 1923; MacGinitie,
1939). These temporary ciliary arrests occur sponta-
neously, in response to general disturbances, or when
undesirable material enters the branchial siphon (Bone
382
CONTROL OF ClONA CILIA
383
and Mackie, 1982), and are usually accompanied by quick
contractions of siphons and mantle ("squirts") (Mac-
Ginitie, 1939). Previous studies on the ascidian branchial
basket showed that ciliary arrests are: ( 1 ) induced by de-
polarizing stimuli, (2) controlled by identified neurons in
the central ganglion, (3) mediated by cholinergic synapses,
(4) correlated with action potentials recorded from the
stigmatal ciliated cells, and (5) dependent on external Ca
(Takahashit'M/., 1973; Mackie etal.. 1974; Arkett, 1987;
Arkett ct ai, 1989). Ascidian stigmatal cilia also undergo
an "inactive" state in which they stand in an upright po-
sition before resuming normal beating (Takahashi et al.,
1973; Mackie etal.. 1974; Arkett, 1987).
Despite the evidence suggesting that ascidian cilia are
controlled by mechanisms similar to those operating in
other systems, neither Ca dependence of arrest or inac-
tivation responses, nor the possible role of cyclic nucleo-
tides in ciliary activity has been investigated directly in
any tunicate.
We have addressed these questions in the ascidian
dona intestinalis by devising an improved method for
microscopic imaging of motile responses of living cilia,
and by developing the first detergent-extracted, ATP-
reactivated cell models of stigmatal cilia. These new ad-
vances enable direct tests of the ionic and biochemical
basis of control of ciliary motility in Ciona and are valu-
able complements to other studies on regulation of sperm
flagellar motility in Ciona (Brokaw, 1987). A preliminary
report of this work has appeared (Bergles and Tamm,
1989).
Materials and Methods
Organism
Specimens of Ciona intestinalis (5-10 cm long) were
obtained from Marine Resources at the Marine Biological
Laboratory and kept in baskets immersed in running sea-
water, or were simply removed from the sides of laboratory
sea-tables in which they had settled and grown. Ciona
specimens were used within a week after removal from
their substrate, because the branchial basket often dete-
riorated after the animals were detached.
Perfusion slides of living stigmata! cilia
Ciona were pinned laterally against a Sylgard-coated
dish, and the tunic was sliced open and pulled back. The
branchial cavity was opened by cutting longitudinally
along the length of the endostyle, then across between the
two siphons. The branchial basket was removed by cutting
the many trabeculae connecting the basket to the mantle.
The excised branchial basket was pinned flat in a Sylgard-
lined petri dish of normal seawater and kept on ice at
0°C. Prior to experimentation, the branchial basket was
further subdivided into small rectangular pieces, about 1
X 5 mm, by cutting between the longitudinal bars with
fine iridectomy scissors. Tissue was used within 2-4 h
after its removal from the animal.
In the final preparation, a piece of branchial basket was
pipetted onto a microscope slide that had been ringed by
a 15 X 50 mm rectangular ridge of petroleum jelly (Vase-
line). The tissue was stretched out near the center of this
rectangular well with the pharyngeal side facing upward.
The ends of the piece were then pressed down against the
slide by fine stainless steel pins anchored in dabs of Vase-
line that had previously been placed on the dry slide. A
square coverslip was placed over the tissue, mounted near
the center of the rectangular well, leaving room on either
side of the coverslip to add and withdraw solution from
the well during perfusion.
Cell models
Small rectangular pieces of branchial basket were placed
in 0.1% saponin, 1% DMSO, 20 mAI EGTA, 150 m.\/
KC1, 10 mAI MgCl:, 30 mAI PIPES, pH 7 (extraction
solution, ES) in a glass well for 4-9 min at room tem-
perature. Tissue was transferred to a second well contain-
ing 2 mAI ATP, 1 mM DTT, 10 mM EGTA, 150 mM
KC1, 10 mAI MgCl:. 30 mM PIPES, pH 7 (reactivation
solution, RS). or to the same solution without ATP (wash
solution. WS). The tissue pieces were then mounted on
perfusion slides for observations. Extraction and reacti-
vation of living pieces mounted on perfusion slides was
not feasible due to distortion of the tissue by muscular
contractions induced by ES.
The effects of ions or reagents on ciliary reactivation
were estimated as follows (Table I). Twenty-five to 100
stigmata with clearly observable cilia were observed for
each tissue piece. A rating of + + + + indicates that >95%
of the cilia in each stigma were beating vigorously. Ratings
of + + + , + + . and + indicate that 75-95%, 50-75%., and
15-50% of the cilia in each stigma were beating, respec-
tively, typically at decreasing frequencies. A rating of
± indicates that 5-15% of the cilia were active, and a
negative rating indicates that <5% were beating. All treat-
ments were repeated on at least five different preparations
of branchial basket pieces.
Reagents and solutions
Calmidazolium and norepinephrine were obtained
from Calbiochem-Behring Corp. (San Diego, California):
sodium metavanadate was obtained from Mallinckrodt
Inc. (St. Louis, Missouri). Calmodulin (bovine brain),
cAMP (bovine brain), protein kinase catalytic subunit
(bovine heart), and all other chemicals were obtained from
Sigma Chemical Co. (St. Louis, Missouri). Ca-EGTA
384
D. BERGLES AND S. TAMM
B
BEATING
ARREST
INACTIVE
QUIESCENT
Figure 1A-C. Diagram of three different states of Ciomi stigmatal cilia, as viewed in cross section of a
stigma to show ciliary profiles. For clarity, only one of the cilia arising from the seven ciliated cells is shown.
The branchial cavity (P) is to the left, the atrial (cloacal) cavity (A) is to the right in all figures. Based on our
observations and Takahashi el al. (1973), Mackie el al. (1974), and Arkett ( 1987). (A) Active cilia beat with
the effective stroke directed towards the atrial chamber (arrows), propelling water out of the pharynx. The
recovery stroke occurs in three dimensions, out of the plane of the power stroke. Metachronal waves (not
shown) travel at right angles to the effective stroke. (B) Arrested cilia lie flat against the stigmatal walls,
inclined in a posture beyond the end of the recovery stroke (stigmata open). (C) Following arrest, cilia stand
upright in an inactive posture (stigma closed) before active beating resumes. Quiescent cilia also remain in
this position.
buffers were prepared according to Salmon and Segall
(1980).
Light microscopy
Perfusion slides were viewed with Zeiss brightfield or
phase-contrast optics ( 16X/0.40 NA or 40X/0.75 NA ob-
jectives), and images were recorded with a DAGE 67M
video camera (Dage-MTI, Michigan City, Indiana 46360)
on a VHS videocasette recorder allowing still-field play-
back (GYYR model 2051, Anaheim, California 92802).
Beat frequency was determined by repetitive counting of
the number of video fields (1/60 s) per beat cycle. Pho-
tographs of still-fields from a video monitor were taken
with an Olympus OM-2N camera on Kodak Tech Pan
(2415)35 mm film.
Results
Stigmatal (•///(,']• system
The anatomy and motility of the stigmatal ciliary sys-
tem of dona and other ascidians have been described
previously (MacGinitie. 1939; Takahashi et al.. 1973;
Mackie et al.. 1974; Arkett. 1987), and are reviewed briefly
here.
The branchial slits or stigmata are lined with seven
rows of laterally flattened cells, each bearing a single
row of cilia (Fig. 1). The ciliated cells are stacked side-
by-side in clusters that abut end-to-end, forming a con-
tinuous ciliated band around the inside of a stigma.
Neurons run within the blood sinus and make synaptic
contacts onto the bases of the ciliated cells, which are
coupled by gap junctions (Mackie el al., 1974; Arkett
et ul.. 1989).
The stigmatal cilia beat outward from the branchial
cavity towards the atrial cavity, generating a water current
that enters the branchial (incurrent) siphon, passes
through the branchial basket, and flows out the atrial si-
phon (Fig. 1A). Ciliary beating is coordinated into dex-
ioplectic metachronal waves that travel unidirectionally
around the stigmatal openings (Fig. 2A).
CONTROL OF CIOKA CILIA
385
In response to mechanical, electrical, or chemical stim-
ulation, all the cilia lining a stigma perform a single rapid
reverse stroke and lie flat against the stigmatal walls for
1-2 s in an arrest position inclined beyond the end of the
normal recover, stroke (Fig. IB). Ciliary arrest halts water
flow into the pharynx and leaves the stigmata completely
open, allowing muscular contractions to "squirt" water
out of the branchial siphon.
Following an arrest, the cilia gradually rise to an upright
position, closing the stigmatal opening (Fig. 1C). The cilia
remain in this straight "inactive" state for a few seconds
before beating resumes and normal metachrony is re-es-
tablished.
In both excised pieces of branchial basket and exposed
intact baskets, some stigmata are always observed with
cilia that stand upright in an "inactive" posture for long
periods. It is not known whether this long-lasting inactive
state was, in any case, preceded by an arrest; but a direct
transition from beating to the inactive position has never
been reported (Takahashieffl/., 1973; Mackie et a!.. 1974:
Arkett, 1987). Upon stimulation, inactive cilia as well as
beating cilia perform an arrest response together (Taka-
hashi et ui. 1973: Mackie et a/.. 1974).
Effects of calcium ionophore
Perfusion of pieces of branchial basket with 100 nM
A23187 in normal seawater for 15-30 min had no no-
ticeable effect on stigmatal ciliary activity. Addition of 50
m\l CaCl: to both the bath and the ionophore suspension
caused most of the cilia to assume an upright inactive
position (stigmata closed) within 5 s after perfusion of
ionophore. Cilia remained in this posture for as long as
observed (up to 5 min). The addition of 100 m.U CaCl:
to the bath and ionophore suspension resulted in stigma-
wide ciliary arrests throughout the field of view within 5
s after perfusion of A23187 (Fig. 2). Ciliary arrests lasted
less than 5- 1 0 s, after which the cilia moved to the inactive
position and remained upright for 15-30 s before resum-
ing beating. When larger pieces of branchial basket were
used, arrests were accompanied by vigorous muscular
contractions.
Perfusion of normal seawater containing 50 mAI or
100 m.U Ca without A23187 but with solvent (0.1%
ethanol/DMSO) did not elicit arrests or inactivations.
Membrane-permeant cAMP analogues
To investigate the possible role of cAMP-regulated
processes (i.e.. activation of PKA) in stigmatal ciliary re-
sponses, we applied membrane-permeant cAMP ana-
logues to branchial basket pieces on perfusion slides.
We directed our attention to stigmata where most of
the cilia were standing upright in a long-lasting inactive
state. Perfusion of 1-10 mM N6-benzoyl-cAMP (B-
Figure 2. Ca icnophore-induced arrest of stigmatal cilia. (A) In sea-
water, metachronal waves of ciliary activity are evident on the stigmatal
wall (arrowhead). (B) Perfusion of A23187 in seawater + 100 m.U Ca
causes ciliary arrest (arrowhead). Video prints; scale bar, 20 Mm.
cAMP) or N6-monobutyryl-cAMP (M-cAMP) in normal
seawater caused many of the quiescent cilia to beat with
normal metachronal coordination within 1-3 min. In
most cases, more than 40% of the cilia became active. A
greater number of quiescent cilia became active in stig-
mata where some of the cilia were already beating prior
to perfusion with B-cAMP or M-cAMP. No significant
activation of cilia was observed after perfusion of the sol-
vent carrier (0.1% ethanol) in seawater without cAMP
analogues.
Cell models
Extraction and reactivation. Treatment of pieces of
branchial basket in ES for 4-9 min stopped most stigmatal
cilia in a more or less upright position. The cilia projected
as tufts from clusters of swollen stigmatal cells, and there
were gaps between the ciliary tufts of adjacent cell groups.
Transfer of tissue to WS did not activate beating; the cilia
remained in a relatively upright posture (Fig. 3A). Shorter
extraction times (2-3 min) resulted in very slow ciliary
beating (<2 Hz) in WS; the cilia usually stopped in the
inactive position within 5-10 min. Thin-section electron
microscopy of tissue extracted for 7 min showed partial
386
D BERGLES AND S. TAMM
or complete removal of ciliary membranes from most of
the stigmatal cilia, while non-extracted tissue prepared by
the same procedure had intact ciliary membranes (not
shown).
Transfer of extracted branchial basket tissue to RS re-
sulted in vigorous beating of 50-70% of the stigmatal cilia.
Ciliary reactivation typically lasted more than 45 min.
Beating sometimes dislodged or displaced the ciliated cells
from the fragile stigmatal wall, causing them to "swim"
through the solution. Normal metachronal waves were
not present. Separated tufts of cilia often beat indepen-
dently, and displayed unicellular metachrony as reported
for reactivated lateral cilia on separated cell groups of
Modiohts demissus gills (Child and Tamm. 1963).
Reactivated beating was also observed by perfusing RS
through a slide of extracted branchial basket in WS (Fig.
3B). RS perfusion caused the cilia to beat hesitantly at
first with a restricted range of motion, then rapidly and
fully within 10-20 s. Reactivated cilia often reached a
steady-state frequency similar to that of living cells ( 10-
14 Hz). Long stretches of cilia sometimes beat nearly syn-
chronously to form common wavefronts (Fig. 3B).
I aihidalc inliihition. RS + 20 fiM vanadate. a potent
inhibitor of dynein ATPase (Gibbons el a/., 1978), did
not reactivate ciliary beating. Norepinephrine (5 mA/) in
RS + vanadate restored reactivated beating, reversing
vanadate inhibition of motility (Table I).
Ca sensitivity. ES-treated pieces of branchial basket
showed normal reactivation of ciliary beating when placed
in wells of 10 7 to 10 b M free Ca (Ca-EGTA buffer).
However, RS + 10~5 to 10 3 M Ca caused the majority
of cilia to assume an upright inactive posture, but never
an arrest position (Table I). A gradual decrease in Ca sen-
sitivity for eliciting inactivation was observed with longer
times in ES; extraction times of more than 9 min often
yielded cilia that did not inactivate in response to Ca.
To check whether a transient Ca-induced arrest re-
sponse might have been missed in depression wells, RS
+ 10 4to 10 3 A/Ca was perfused into a slide containing
a piece of extracted branchial basket in WS. No momen-
tary arrest response was observed before the cilia assumed
a rigidly straight inactive position.
Various approaches were tried to elicit arrests in RS
+ 10 5 to 10 3 A/Ca. For example, a cocktail of protease
inhibitors ( 1 mg/ml trypsin inhibitor; 0.5 mg/ml leupep-
tin: I mA/ TAME; 0.2 mg/ml PMSF. 1 mg/ml BSA) were
included in both ES and RS to prevent possible proteolysis
of putative Ca sensors or proteins required to mediate
arrest. Different detergents (Brij-5 8, Brij-35, Triton-X 100)
were tried in place of saponin to preclude the extraction
of Ca-binding proteins (i.e.. calmodulin). In another series
of experiments, K acetate was substituted for KC1 in both
ES and RS. None ot these modifications resulted in ciliary
arrests in RS + 10 J M Ca. We did note, however, that
reactivation was consistently better in solutions containing
KC1 rather than K acetate (Table I).
Calinociiilin. Addition of calmodulin antagonists, 100
\iM trifluoperazine (TFP) or 100 nAl calmidazolium to
RS + 50-100 fiAlCa restored reactivated beating, thereby
reversing Ca-dependent inactivation (Table I). TFP typ-
ically gave more consistent results than did calmidazo-
lium. Chlorpromazine (100 ^M) did not significantly re-
verse the inactivation of cilia in RS + Ca. These results
indicate that calmodulin mediates Ca-dependent inacti-
vation of stigmatal cilia.
Addition of 65 /ug/ml bovine brain calmodulin to RS
+ 100 n.M Ca did not elicit an arrest of cilia in tissue
initially bathed in WS or RS (Table I).
cAMP and PK.4
RS + 100 nM cAMP and 1 m.U IBMX. a cyclic nu-
cleotide phosphodiesterase inhibitor substantially in-
creased both the proportion and vigor of ciliary reacti-
vation compared to tissue incubated in RS alone (Table
I). Thus cAMP and Ca exert opposing effects on ciliary
reactivation. In RS containing 100 nM cAMP + 100 nM
Ca, cilia vibrated or twitched rapidly in a rigid inactive
position. Addition of 28 Mg/ml catalytic subunit of PK.A
to RS likewise improved the extent of ciliary reactivation
to more than 95% in most cases (Table I).
Discussion
Calcium and arrest
The role of Ca in triggering a variety of ciliary and
flagellar motor responses is well documented (Eckert and
Murakami, 1972; Naitoh and Kaneko, 1972; Tsuchiya,
1977; Hyams and Borisy. 1978; Walter and Satir. 1978;
Gibbons and Gibbons, 1980; Brokaw and Nagayama,
1 985; Nakamura and Tamm, 1985; Satir, 1985;Stommel
and Stephens. 1985: Machemer, 1986: Brokaw. 1987,
1991; Tamm. 1988; Otter, 1989).
Although arrest of stigmatal cilia in ascidians has long
been suspected to be Ca-dependent (Takahashi el a/..
1973; Mackie el ai, 1974). direct evidence for this has
been lacking. Our finding that Ca ionophore in the pres-
ence of 100 m.U Ca elicits arrest of dona stigmatal cilia
strongly argues for the Ca-dependency of this response.
However, these experiments were performed on pieces of
branchial basket tissue, and stigmatal cilia have been
shown to be under neuronal control (Mackie et a/.. 1974;
Arkett, 1987). Therefore, our results could also be ex-
plained by ionophore-mediated influx of Ca at presynaptic
sites mediating nervous control of ciliary arrest, without
requiring Ca influx into the ciliated cells themselves.
To directly test whether ciliary motility in ascidians is
regulated by Ca, we prepared the first ATP-reactivated
CONTROL OF CIONA CILIA
387
IB.
Figure 3. Stigmatal cell models. Cilia lining a stigma are shown on the left and diagrammed on the
right. (A) A single stigma in WS. Tufts of immotile cilia (black) stand upright, projecting from the refractile
wall of the stigma into the Stigmatal space. In the diagram, the wall (edge) of the stigma is indicated by the
irregular horizontal lines. (B) The same stigma after perfusion of RS. The cilia beat vigorously and, to a
large extent, synchronously, giving rise to common wavefronts (arrowheads). As a result, the stigma is more
open than in A. The Stigmatal wall does not change (compare outlines of stigma in diagrams). Scale bar. 1(1
fim. A. B
models of Stigmatal ciliated cells. We were unable to elicit
ciliary arrest in our saponin-permeabilized models at any
Ca concentration used (10 5 to 10 \UCain RS). Instead,
the axonemes stopped in an upright inactive position
without passing through an arrest.
We were concerned that our permeabilization proce-
dure, or possibly subsequent proteolysis, might have re-
moved or destroyed critical control factors or Ca-binding
proteins (i.e.. calmodulin) necessary for demonstrating
Ca-sensitivity of arrest in cell models. For example, ex-
traction of calmodulin from sea urchin sperm and protist
cilia leads to modification or loss of Ca control of axo-
nemal motor responses (Brokaw and Nagayama, 1985;
Izumi and Miki-Noumura. 1985; Izumi and Nakaoka,
1987). Extraction and incubation procedures may also
modify the calmodulin-binding affinity of the axoneme
(Brokaw. 1991 ). In addition, some detergents commonly
used to make cell models (i.e., Triton-X 100) are potent
inhibitors of both calmodulin and calmodulin-dependent
cyclic nucleotide phosphodiesterase (Sharma and Wang,
1981).
However, our attempts to restore presumed Ca sensi-
tivity of ciliary arrest in models by trying different deter-
gents, extraction times, protease inhibitors, or addition of
exogenous bovine brain calmodulin, were uniformly un-
successful. Nevertheless, the variable sensitivity of the in-
activation response of models to Ca, particularly after
longer extraction times (more than 9 min in ES), suggests
that the absence of Ca-sensitive arrest in our models may
be due to loss or modification of an as yet unidentified
factor.
Calcium and inaclivalion
The upright inactive posture of Stigmatal cilia is clearly
Ca-dependent: reactivated cilia are inactivated by 10~5 to
10 3 M Ca, and cilia on living tissue are inactivated by
Ca ionophore in the presence of a lower Ca concentration
(50 mA/) than that leading to arrest ( 100 mM. see above).
These findings suggest that inactivation has a lower
threshold to intracellular Ca than does arrest.
Inactive cilia of cell models in RS -f- Ca were induced
to beat by the addition of a calmodulin antagonist, either
388
D. BERGLES AND S. TAMM
Table I
/:"//( 'i 7 v ni nirn>n\ compounds un rcactivatum of ciliary motility
in < c
Solution
Ciliary
activity
WS
RS(KCI)* ++( + )
RS(K. Acetate) + +
RS + 20 fi.\l Vanadate
RS + 20 nM Vanadate + 5 m.\I Norepinephnne
RS + 10//;UCa:+
RS + 50
RS+
RS + lOOfiA/Ca2* + Pis
RS + 100 fiMCa2+ + 100 pM Trifluoperazine
RS + 100 nM Ca:+ + 100 M.U Calmidazolium
RS + 100 nM Ca2+ + 100 rfl Chlorpromazine
RS + 100 n.M Ca:+ + Calmodulin (65 Mg/ml)
RS + 100 /u.UcAMP + 1 mM IBMX
RS + 100 11 M Ca:+ + 100 pM cAMP
RS + PKAcs (28
+ + +
+
±
+ + +
++
++++
- vibrating
+ + + +
* Standard RS used below.
Plus and minus ratings indicate relative degrees of ciliary activity (see
Materials and Methods); minus indicates that nonbeating cilia were in
an upright inactive position.
Abbreviations: WS, wash solution; RS, reactivation solution; IBMX,
isobutylmethyKanthine; Pis. protease inhibitors; PKAcs. catalytic subunit
of protein kinase.
TFP or calmidazolium. This suggests that Ca-induced in-
activation of dona stigmatal cilia is mediated by cal-
modulin. Other Ca-dcpcndcnt ciliary motor responses,
such as arrest of mussel gill lateral cilia (Reed el a/.. 1982;
Stommel, 1984), activation of Mytilux gill abfrontal cilia
(Stommel, 1984), and reorientation of ciliary beat direc-
tion in some cell models of Pcirdiuec/uni (Otter ct til..
1984; Izumi and Nakaoka, 1987) and Tetrahymcna
(Izumi and Miki-Noumura, 1985), are also partially or
completely inhibited by anti-calmodulin drugs.
Ca may exert its effects on ciliary motility by activating
calmodulin-dependent protein kinase (C kinase) or phos-
phatase (calcineurin), thus changing the phosphorylation
levels of axonemal regulatory proteins (Nakaoka and Ooi,
1985; Tash, 1989; Hamasaki ct til.. 1989; Bonini ct til..
1991). Because the catalytic subunit of PICA did not in-
activate reactivated dona stigmatal cilia, but rather en-
hanced motility, the mechanism by which Ca-calmodulin
is presumed to inactivate stigmatal cilia may involve a
dephosphorylation reaction.
The transient inactive state exhibited by dona cilia
after every arrest resembles the transient inactivation of
Paramecium cilia that occurs after depolarization-induced
Ca-dependent reversal of beat direction, before the beat
cycle is renormalized (Machemer, 1986). Voltage-clamp
experiments with Puninwciwn showed that an inactive
state, or frequency minimum, also intervenes between
normal beating and the onset of stimulus-induced ciliary
reversal. In Paramecium, a transient inactivation response
therefore precedes and follows the Ca-dependent ciliary
reversal response, suggesting that inactivation may be
caused by a Ca concentration slightly elevated above nor-
mal resting level (Machemer, 1986). The transient inactive
state following the arrest of dona cilia may also reflect
an intermediate level of internal Ca concentration. The
epaulette cilia of echinoplutei larvae sometimes undergo
a similar upright inactive state after Ca-dependent reversed
beating (Mogami ct a/.. 1991).
Finally, a recent study of Ca-induced asymmetry of
ATP-reactivated flagellar bending waves of sea urchin
sperm indicates the existence of two separate Ca responses,
mediated by high-affinity and lower-affinity Ca sensors
(Brokaw. 1991).
cAMP and quiescence
Cyclic nucleotides (cAMP, cGMP) also play a role in
regulating ciliary and flagellar motility: in particular.
cAMP is typically involved in initiating and maintaining
ciliary and flagellar beating (Opresko and Brokaw, 1983;
Stommel and Stephens. 1985; Takahashi ct ul.. 1985:
Murofushi ct ul.. 1986; Brokaw, 1987; Murakami,
1987a,b: Stephens and Stommel. 1989; Tash. 1989; Bo-
nini ci til.. 1991).
Neuronal activation of quiescent lateral cilia on Mytilus
gill is due to 5 HT-triggered augmentation of cellular
cAMP levels, leading to cAMP-dependent protein kinase-
mediated phosphorylation of axonemal dynein light
chains (Stephens and Prior. 1990). Quiescence of lateral
cilia is thus believed to reflect lowered cAMP concentra-
tion and resultant dephosphorylation of dynein polypep-
tides. Phosphorylation of axonemal dynein polypeptides
has also been reported in other systems (Hamasaki and
Satir, 1989; Chilcote and Johnson. 1990; Dey and Brokaw
1991; Stephens and Prior, 1 99 1 ).
We found that membrane-permeant cAMP analogs
stimulate the beating of Cioiui stigmatal cilia held in a
long-lasting inactive state (termed quiescence). Moreover,
reactivated beating of stigmatal ciliary models is improved
by cAMP and IBMX, or the addition of the catalytic sub-
unit of protein kinase to RS. Reactivation of Mytilus lat-
eral ciliary models is also improved by the presence of
the catalytic subunit. which can override Ca arrest (Stom-
mel, 1984; Stommel and Stephens, 1985; Stephens and
Stommel, 1989). These findings suggest that the quies-
cence (long-lasting inactivation) of dona cilia is physio-
logically similar to the quiescence ofMyti/ns lateral cilia;
i.e.. that increased cAMP levels may also be responsible
for maintaining the activity of dona stigmatal cilia via
cAMP-dependent phosphorylation of regulatory axone-
mal polypeptides.
CONTROL OF CIONA CILIA
389
Although their underlying biochemical mechanisms
seem to be similar, the postures of quiescent Mylilus lateral
cilia and quiescent Ciona sigmatal cilia are quite different.
Lateral cilia rest at the end of the recovery stroke, whereas
stigmatal cilia stand upright, midway between the effective
and recovery strokes.
Ciona cilia thus remain upright during both the tran-
sient inactive state and the quiescent or long-lasting in-
active state. However, quiescence differs from inactiva-
tion. not only by its longer duration, but also by the ap-
parent absence of a preceding arrest.
cAMP and Ca act antagonistically on Mytilus lateral
cilia, as well as on several other ciliary and flagellar systems
(see Brokaw, 1987: Stephens and Stommel. 1989; Bonini
el al.. 1991). Moreover. cAMP or cAMP-dependent pro-
tein kinase can override the Ca effect and activate beating
of Ca-arrested cilia (Murakami and Takahashi, 1975:
Murakami. 1983: Stommel and Stephens, 1985). Our
finding, that adding Ca and cAMP to RS causes rapid
vibration or twitching of stigmatal cilia in the straight
inactive position, indicates a similar antagonism between
cAMP-mediated activation of beating and Ca-induced
inactivation (rather than arrest) of cilia.
In conclusion, we have developed an in vitro prepara-
tion of Ciona branchial basket cilia which allows inves-
tigation for the first time of the ionic and molecular control
of ciliary motility in tunicates. Further studies using im-
proved models should provide more detailed and quan-
titative information for comparison to other systems.
Acknowledgments
We thank Dr. Ray Stephens, MBL, for helpful discus-
sions and advice, and Signhild Tamm for electron mi-
croscopy. Dorothy Hahn patiently and skillfully processed
these words. This research fulfilled the requirements for
Independent Work of Distinction by DB at the Boston
University Marine Program, and was supported by NIH
grant GM 27903 to SLT.
Literature Cited
Arkett, S. A. 1987. Ciliary arrest controlled by identified central neurons
in a urochordate (Ascidiacea). J Comp. Physiol A 161: 837-847.
Arkett, S. A., G. O. Mackie, and C. L. Singla. 1989. Neuronal orga-
nization of the ascidian (LIrochordata) branchial basket revealed by
cholinesterase activity. Cell Tiss. Res 257: 285-294.
Bergles, D. E., and S. L. Tamm. 1989. Calcium-sensitive ATP-reac-
tivated models ofCuma intestinalis branchial basket cilia. Bioi Bull-
Ill: 313.
Bessen, M., R. B. Fay, and G. B. \\itman. 1980. Calcium control of
waveform in isolated flagellar axonemes of Chlamydomonas. J. Cell
Bail. 86: 446-455.
Bone, Q., and G. O. Mackie. 1982. LIrochordata. Pp. 473-535 in Elec-
trical Conduction and Behaviour in 'Simple' Invertebrates. G. A. B.
Shelton. ed. Oxford University Press. Oxford. U.K.
Bonini, N. M., T. C. Evans, L. A. P. Miglietta, and D. L. Nelson.
1991. The regulation of ciliary motility in Parameciiim by Ca2+
and cyclic nucleotides. Pp. 227-272 in Advances in Second Messenger
and Phosphopmiein Research, vol 23. P. Greengard and G. A. Rob-
ison. eds. Raven Press, New York.
Brokaw, C. J. 1979. Calcium-induced asymmetrical beating of Triton-
demembranated sea urchin sperm flagella. J Cell Biol. 82: 401-41 1.
Brokaw, C. J. 1987. Regulation of sperm flagellar motility by calcium
and cAMP-dependent phosphorylation. / Cell Biochem. 35: 175-
184.
Brokaw, C. J. 1991. Calcium sensors in sea urchin sperm flagella. Cell
Moid. Cywskei 18: 123-130.
Brokaw, C. J., R. Josslin, and L. Bobrow. 1974. Calcium ion regulation
of flagellar beat symmetry in reactivated sea urchin spermatozoa.
Biochem. Biopln-s. Re*. Commim. 58: 795-800.
Brokaw, C. J., and S. M. Nagayama. 1985. Modulation of the asym-
metry of sea urchin sperm flagellar by calmodulin. / Cell Biol 111(1:
1875-1883.
Carter, G. S. 1926. On the nervous control of the velar cilia of the
nudibranch veliger. / Exp. Biol. 4: 1-26.
Chilcote, T. J., and K. A. Johnson. 1990. Phosphorylation of Tetra-
hymena 22 S dynem. ./ Biol. Chem 265: 17257-17266.
Child, F. M., and S. I,. Tamm. 1963. Metachronal ciliary coordination
in ATP-reactivated models of Modiolus gills. Biol Bull 25: 373-
374.
Dey, C. S., and C. J. Brokaw. 1991. Activation of Ciona sperm motility:
phosphorylation of dynein polypeptides and effects of a tyrosine kinase
inhibitor./ Cell Sci. 100: 815-824.
Eckert, R., and A. Murakami. 1972. Calcium dependence of ciliary
activity in the oviduct of the salamander Nectiirus. J. Physiol. 226:
699-711.
Eckert, R., V. Naitoh, and H. Machemer. 1976. Calcium in the bio-
electric and motor functions of Parameciiim. Symp. Soc. Exp. Biol.
30: 233-255.
Fedele, M. 1923. Le attivita dinamiche ed i rapporti nervosi nella vita
dei Dolioli. Puhl. Sla:. Zoo/. Napoli 4: 129-239.
Gibbons, I. R., M. P. Cosson, J. A. Evans, B. H. Gibbons. B. Houck.
K. H. Martinson. \V. S. Sale, and \\.-J. V. Tang. 1978. Potent
inhibition of dynein adenosine tnphosphatase and of the motility ot
cilia and sperm flagella by vanadate. Proc. Nail. Acad. Sci. USA 75:
2220-2224.
Gibbons, B. H., and I. R. Gibbons. 1980. Calcium-induced quiescence
in reactivated sea urchin sperm. J Cell Biol 84: 13-27.
Hamasaki, T., T. J. Murtaugh, B. H. Satir, and P. Satir. 1989. In vitro
phosphorylation of Parameciiim axonemes and permeabilized cells.
CellMotil. Cytoskel. 12: 1-11.
Hamasaki, T., and P. Satir. 1989. Parameciiim dynein contains a
phosphorylatable 29 kDa light chain. CellMotil. Cytoskel. 11: I89A.
Holwill, M. E. J., and J. L. McGregor. 1976. Effects of calcium, on
flagellar movement in the trypanosome Crithidia oncopelti J. Exp.
Biol. 65: 229-242.
Hyams. J. S., and G. G. Borisy. 1978. Isolated flagellar apparatus of
Chlamydomonas: characterization of forward swimming and alter-
ation of waveform and reversal of motion by calcium ions in vitro.
J. Cell Sci 33:235-253.
Izumi, A., and T. Miki-Noumura. 1985. Tetrahymena cell model ex-
hibiting Ca-dependent behavior. Cell Motil 5: 323-331.
Izumi, A., and V. Nakaoka. 1987. cAMP-mediated inhibitory effect of
calmodulin antagonists on ciliary reversal of Parameciiim. CellMotil.
1: 154-159.
MacGinitie, G. E. 1939. The method of feeding of tunicates. Biol
Bull 77: 443-447.
Machemer, H. 1986. Electromotor coupling in cilia. Pp. 205-250 in
Membrane Control ol Cellular Activity. H. C. Luttgau. ed. Gustav
Fischer Verlag. New York.
390
D. BERGLES AND S. TAMM
Mackie, G. O., D. H. Paul, C. M. Singla, M. A. Sleigh, and D. E.
Williams. 197-4. Branchial innervation and ciliary control in the
ascidian ( <> <c. R Soc Lond. B 187: 1-35.
Mogami, V., K. Fisjima, and S. A. Baba. 1991. Five different states of
ciliary activity in the epaulette of echinoplutei. ./ Exp. Binl. 155:
65-75.
Murakami, A. 1983. Control of ciliary beat frequency in Mytilus J
Siibmicrosc. Cylol 15: 313-316.
Murakami. A. 1987a. Control of ciliary1 beat frequency in the gill of
Mytilus. 1. Activation of the lateral cilia by cyclic AMP. Comp.
Biochem. Physiol 86C: 273-279.
Murakami, A. 1987b. Control of ciliary beat frequency in the gill of
Mytilus. II. Effects of saponin and Brij-58 on the lateral cilia. Comp
Biochem Physiol 86C: 281-287.
Murakami, A., and k. Takahashi. 1975. The role of calcium in the
control of ciliary movement in Mytilttx. II. The effects of calcium
ionophores x537A and A23187 on the lateral gill cilia. / Fac. Sci.
Univ. Tokyo 11' 13: 251-256.
Murofushi, II., k. Ishiguro, D. Takahashi. J. Ikeda, and H. Sakai.
1986. Regulation of sperm flagellar movement by protein phos-
phorylation and dephosphorylation. Cell Motil. 6: 83-88.
Naitoh, V., and H. Kancko. 1972. Reactivated Triton-extracted models
of Paramecium. modification of ciliary movement by calcium ions.
Science 176: 523-524.
Nakamura, S., and S. L. Tamm. 1985. Calcium control of ciliary re-
versal in ionophore-treated and ATP-reactivated comb plates of
ctenophores. ./. Cell Biol. 100: 1447-1454.
Nakaoka, Y., and II. Ooi. 1985. Regulation of ciliary reversal in Tnton-
extracted Paruinecium by calcium and cyclic adenosine monophos-
phate. / CeltSci. 77: 185-195.
Opresko, L. K., and C. J. Brokaw. 1983. cAMP-dependent phosphor-
ylation associated with activation of motility oft'ionu sperm flagella.
Gamete Rex. 8: 201-218.
Otter. T. 1989. Calmodulin and the control of flagellar movement.
Pp. 281-298 in Cell Movement, vol I The Dynein ATPases. F. D.
Warner. P. Satir. and I. R. Gibbons, eds. Alan R. Liss, New York.
Otter, T., B. Satir, and P. Satir. 1984. Tntluoperazine-induced changes
in swimming behavior of Paramecium: evidence for two sites of drug
action. Cell Molil. 4: 249-267.
Preston, R. R., and Y. Saimi. 1990. Calcium ions and the regulation
of motility in Paramecium. Pp. 173-200 in Ciliary and Flagellar
Membranes. R. A. Bloodgood, ed. Plenum. New York.
Reed, \Y.. S. Lebduska. and P. Satir. 1982. Effects of trifluoperazine
upon the calcium dependent ciliary arrest response of freshwater
mussel gill lateral cells. Cell Motil. 2: 405-427.
Salmon, K. D., and R. R. Segall. 1980. Calcium-labile mitotic spindles
isolated from sea urchin eggs (Lytechimis variegutus). J. Cell Biol.
86: 355-365.
Satir, P. 1985. Switching mechanisms in the control of ciliary motility.
Mud. Cell Biol 4: 1-46.
Sharma, R. K., and J. H. Wang. 1981. Inhibition of calmodulin-ac-
tivated cyclic nucleotide phosphodiesterase by Triton X-100.
Biochem. Biophys Res Commiin 100: 710-715.
Stephens, R. E., and E. W. Stommel. 1989. Role of cyclic adenosine
monophosphate in ciliary and flagellar motility. Pp. 299-316 in Cell
Miiveme/it. vol I The Dynein ATPases. F. D. Warner. P. Satir. and
I. R. Gibbons, eds. Alan R. Liss, New York.
Stephens, R. E., and G. Prior. 1990. Cyclic AMP-dependent dynein
light chain phosphorylation in neuronally-controlled mussel gill cilia.
/ Cell Biol 111: 295a.
Stephens, R. E., and G. Prior. 1991. Cyclic AMP-dependent phos-
phorylation of dynein alpha-heavy chains in Mytilus edulis sperm
flagella. Biophys. J. 59: 567A.
Stommel. E. W. 1984. Calcium activation of mussel gill abfrontal cilia.
./ Comp Physiol. A 155:457-469.
Stommel, E. W., and R. K. Stephens. 1985. Cyclic AMP and calcium
in the differential control of .\fytilu.i gill cilia. ./ Comp Physiol A
157: 45|_45M.
Takahashi, D., II. Murofushi, k. Ishiguro, J. Ikeda, and H. Sakai.
1985. Phosphoprotein phosphatase inhibits flagellar movement of
Triton models of sea urchin spermatozoa. Cell Struct. Fund 10:
327-337.
Takahashi. D., S. A. Baba, and A. Murakami. 1973. The "excitable"
cilia of the tunicate Ciima nncslinalis. J. Fac. Sci. I'mv. Tokyo II'
13: 123-137.
I .num. S. I.. 1988. Calcium activation of macrocilia in the ctenophore
Beroe. ./. Comp Physiol. A 163: 23-31.
Tash, J. S. 1989. Protein phosphorylation: the second messenger signal
transducer of flagellar motility. Cell Molil. Cytoskel. 14: 332-339.
Tsuchiya, T. 1977. Effects of calcium ions on Triton-extracted lamel-
libranch gill cilia: ciliary arrest in a model system. Comp. Biochem.
Physiol 56A: 353-361.
Walter, M. F., and P. Satir. 1978. Calcium control of ciliary arrest in
mussel gill cells. / Cell Biol 79: 1 10-120.
Reference: Biol. Bull. 182: 391-400. (June, 1992)
Slow Photic and Chemical Induction
of Bioluminescence in the Midwater Shrimp,
Sergestes similis Hansen
MICHAEL I. LATZ* AND JAMES F. CASE
Department of Biological Sciences and Marine Science Institute, University of California.
Santa Barbara. California 93106
Abstract. The initial luminescent response to photic
stimulation of dark-maintained specimens of the mid-
water shrimp, Sergestes similis Hansen, differed from the
conventional counterillumination response. Animals were
initially unresponsive to light; bioluminescence was only
induced after a latency of 3 min. Maximum intensity was
reached after approximately 25 min. During the induction
process, light emission from the anterior light organs was
frequently observed prior to output from the posterior
organ. Once luminescence was induced, responses exhib-
ited the typical fast kinetics of the counterillumination
response and changes in light organ output occurred syn-
chronously.
Visual input was necessary to maintain this state. Dark
readaptation of counterilluminating animals resulted in
a return to the slow response kinetics characteristic of
untested animals. Because eyestalk ablation or crushing
caused immediate production of luminescence in previ-
ously untested animals, the slow induction did not involve
the ability of the light organs to produce light.
Serotonin was effective in stimulating bioluminescence
in intact animals; the induction of light emission pro-
ceeded at a rate similar to that for photic stimulation.
Other putative neurotransmitters, including norepineph-
rine, acetylcholine, GABA, and L-glutamic acid, did not
stimulate bioluminescence. Isolated light organs exhibited
high background levels of light emission, which were un-
changed by serotonin treatment. However, serotonin was
Received 31 October 1991; accepted 6 March 1992.
* Present address: Marine Biology Research Division 0202, Scnpps
Institution of Oceanography. University of California, San Diego, La
Jolla. California 92093.
Abbreviations: ACh, acetylcholine; GABA. gamma aminobutync acid;
PCA, p-chloroamphetamine; 5-MT. 5-metho\ytryptamme.
effective in stimulating luminescence in animals with
ablated eyestalks. These results suggest a dual control sys-
tem involved in the induction and maintenance of bio-
luminescence in 51. simi/is.
Introduction
Marine organisms are vulnerable to predation by up-
wards-viewing predators that scan for prey silhouetted against
downwelling illumination. In some midwater animals, this
vulnerability may be reduced by luminescent countershad-
ing, or counterillumination, in which downward-directed
bioluminescence replaces oceanic light absorbed or reflected
by the animal's body (Clarke, 1963; Herring. 1982; Young,
1983). For counterillumination to be optimally effective,
light emission must match the spectrum, intensity, and di-
rection of ambient light, so that bioluminescence effectively
replaces ambient downward-directed illumination. Strong
experimental evidence for a counterillumination role of lu-
minescence exists for midwater squids, fishes, and crusta-
ceans (reviewed by Young, 1983).
Bioluminescence by the decapod shrimp, Sergestes
similis Hansen, functions in this manner (Warner et al,
1979), counterilluminating the body by matching the
spectral distribution (Herring, 1983; Widder el al. 1983),
intensity (Warner et al.. 1979), and angular distribution
(Latz and Case, 1982) of oceanic downwelling illumina-
tion. Light emission by S. similis associated with coun-
terillumination is stimulated only by downward-directed
illumination and can be maintained for long periods
(Warner et a/.. 1979). In the dark, no luminescence is
produced for counterillumination.
Little is known of the physiological control of coun-
terillumination. Luminescence by 5. similis is regulated
by visual input; when the eyes are masked, light emission
391
392
M. I. LATZ AND J. F. CASE
is absent. Luminescent response latencies to visual stim-
ulation of only a few seconds are consonant with either
neural or hormonal control (Warner et cil.. 1979). Bio-
luminescence originates from modified portions of the
hepatopancreas, the organs of Pesta (Dennell, 1940; Her-
ring, 1981). The mechanism of control of light emission
by the organs of Pesta is unclear because the light organs
have neither been shown to be innervated nor lumines-
cence to be electrically or chemically excitable (Herring,
1976, 1981).
In other midwater animals, bioluminescence used for
counterillumination appears to be under neural control.
In squids, morphological and physiological evidence sup-
ports direct neural control (Arnold and Young, 1974; Dilly
and Herring, 1974; Herring, 1977). The photophores and
caudal organs of myctophid fishes are under neural con-
trol, even though the chemical basis remains obscure.
They are richly innervated and electrically or neurally
excitable (reviewed by Herring, 1982).
The present study documents a previously undescribed
aspect of counterillumination by S. similis: the slow initial
induction of luminescence in previously untested, dark-
maintained animals, which occurs prior to the counter-
illumination response. This induction can be mimicked
by chemical treatment with the neurotransmitter sero-
tonin. The slow kinetics of photic and chemical induction
compared to the typical counterillumination response
suggest different mechanisms controlling these responses.
Results support the hypothesis that a blood-born factor,
perhaps via a neurosecretory pathway, is involved in the
induction process.
Materials and Methods
Adult specimens of Sergestes similis were collected at
night from depths of 75-200 m in the Santa Barbara Basin,
near Santa Barbara. California, using a midwater trawl.
Trawl contents were recovered under dark conditions on
moonless nights and sorted under dim red light. Animals
were placed in chilled seawater, brought into the labora-
tory within 3 h of collection, and were maintained in 100-
1 aquaria with flow-through, sand-filtered seawater (10°C).
All tests were performed within one week of collection,
during which time animals remained in good physiological
condition and exhibited low mortality. Only actively
swimming specimens were used for testing. Except for
brief exposure to dim red light during handling, animals
remained in constant darkness and were not fed.
For testing, specimens were loosely restrained by a
clamp around the cephalothorax and placed in a sealed,
clear acrylic chamber (1.75 > 2.5 X 10 cm) filled with
10°C seawater (Fig. 1A). Bioluminescence was induced
by downward-directed illumination conducted by a fiber
optic light guide from a tungsten-halogen source (Dolan-
Jenner Inc.) to a 465 nm interference filter (Ditric Optics,
half band width 9.4 nm) and diffused by two opal ground
glass plates. Light intensity was regulated by neutral den-
sity filters (Rolyn Optics) and measured by a LJnited De-
tector Technology Inc. 40X Optometer. Stimulus duration
was controlled by an electro-mechanical shutter (Vincent
Associates) (Fig. 2). Stimulus intensities were comparable
to light intensities of <1 X 10~6 to 5 X 10~2 ^W cm :
present at daytime depths frequented by S. similis in the
Santa Barbara Basin (Clarke, 1966).
Photomultiplier recordings
For these long-term experiments, the seawater in the
acrylic chamber containing the restrained animal was ex-
changed at a rate of approximately 50 ml min'1. The
apparatus for light stimulation was as described above.
Bioluminescence was detected by an EMI 978 IB photo-
multiplier operating at -550 V and fitted with an electro-
mechanical shutter (Fig. 2). The photomultiplier was lo-
cated 10 cm beneath the animal. The stimulus light and
the photomultiplier were isolated by a pair of rotating
light choppers (Rofin) producing 5 ms light pulses at 100
Hz, synchronized 180° out of phase with each other and
positioned one above and one below the experimental
chamber. Consequently, the photomultiplier viewed the
specimen in the dark interval between light pulses deliv-
ered to the specimen. The test animal perceived the light
stimulus as a continuous source, because the chopping
rate was greater than the critical flicker fusion frequency
of marine crustaceans, which is typically below 60 Hz
(Waterman, 1961). The chopped photomultiplier signal
was led through a Keithley 427 Current Amplifier, rec-
tified by a Keithley Autoloc 840 Amplifier referenced to
the chopping frequency, and displayed on a Grass 79D
Polygraph. A photodiode monitored the filtered light
stimulus and registered stimulus presentations on the
polygraph record.
Specimens in the chamber were acclimated in the dark
for at least 20 min following handling under dim red light.
They were then subjected to light stimuli ranging from
2 X 10 5 to 4 X 10~4MWcirr2.
The intensity of bioluminescence was measured from
the polygraph record as amount of baseline shift corrected
for dark current, and expressed as photomultiplier anode
current. The apparatus was not calibrated for luminescent
output in irradiance units.
Image intensificalii >n
Bioluminescence from restrained animals was viewed
from below with an image intensifier (EMI Type 9912,
four-stage, maximum radiant power gain 106 at 440 nm),
fitted with a 75 mm f/1.9 objective lens, by means of a
first-surface mirror positioned beneath the chamber at
BIOLUMINESCENCE IN A SERGESTID SHRIMP
393
Figure 1. Views of intact and dissected preparations of Sergestes similis- (A) Dorsal view of living intact
specimen restrained in testing chamber. Specimen was loosely clamped about the midregion of the body
during experimentation. Immediately anterior to the clamp is the hepatopancreas and tbregut. The chamber
was superfused with chilled ( 10°C) filtered seawater. Scale bar = 5 mm. (B) Ventral view of isolated hepa-
topancreas showing locations of luminous tissue. Dark pigmentation characteristic of the luminous tissue
(arrows) is associated with the (a) anterolateral pair of organs of Pesta. (m) lateral midgastric pair of organs,
and (p) posterior fringe organ. Scale bar = 1 mm.
an angle of 45°. Typical operating voltage was 34 kV.
The anode phosphor of the image intensifier was viewed
by a Panasonic newvicon video camera with a 25 mm
f/0.95 objective lens, and images were recorded on vid-
eotape together with a time and video frame reference.
The apparatus for stimulus illumination was as described
above.
The chamber containing a restrained animal was po-
sitioned in the dark in the experimental apparatus. During
experiments, the stimulus intensity was either 1 X IfT5
or 2 X 10~4 jiW cm"2. At one-minute intervals, the stim-
ulus was briefly extinguished to permit documentation of
bioluminescence.
Chemical stimulation
The physiological basis of the slow photic induction of
bioluminescence was further investigated with tests of
putative invertebrate neurotransmitters. For this study,
specimens of S. similis were collected during the day from
the Santa Barbara Basin and thereafter maintained in
darkness and handled under dim red light. Intact live an-
imals were restrained in the test chamber. In some cases,
the hepatopancreas tissue with attached light organs was
isolated by dissection, pinned in a clear dish layered with
Sylgard, and placed in the test chamber. In some speci-
mens, both eyestalks were ablated at their bases with iri-
dectomy scissors prior to chemical testing. Biolumines-
394
M. I. LATZ AND J. F. CASE
LIGHT SOURCE
FIBER OPTIC
SHUTTER
LIGHT CHOPPER
OPTICAL FILTERS
DIFFUSE RS
SPECIMEN
CHAMBER OUT
LIGHT CHOPPER
SHUTTER
PHOTOMULTIPLIER
STIMULUS
BIOLUMINESCENCE
CHART RECORD
Figure 2. Schematic of experimental apparatus used to measure the
intensity oi bioluminescence during countenllumination. The specimen
in the testing chamber superfused with chilled seawater (SW) was sub-
jected to a diffuse downward-directed illumination of controlled intensity
and wavelength, pulsed at approximately 100 Hz by a light chopper
(dashed lines above specimen). Downward-directed luminescence (solid
lines below specimen) was chopped (dashed lines below specimen) by a
second light chopper, synchronized 180° out of phase with the stimulus
chopper, and was detected by a photomultiplier. The second light chopper
prevented the stimulus illumination from reaching the detector. The
bioluminescence signal was amplified, rectified by a lock-in amplifier,
and displayed on a chart recorder along with a stimulus record obtained
from a photodiode monitoring the stimulus illumination. Not drawn to
scale.
cence was detected from below the chamber by an EMI
970 IB photomultiplier operating at -750 V and fitted
with an electromechanical shutter. The photomultiplier
signal was amplified by a Keithley 427 Current Amplifier
and displayed on a Grass 79D Polygraph. Levels of light
emission were expressed as PMT anode current, without
radiometric calibration.
The action of neurotransmitters was assayed with intact
specimens or isolated hepatopancreas tissue containing
the organs of Pesta. The following solutions were prepared
in filtered seawater: 1 X 10~3 A/ acetylcholine (ACh), 1
X 10~3 M gamma aminobutyric acid (GABA), 1 X 10~3
M L-glutamic acid, 1 X 10~3 M norepinephrine, and 5.7
X 10~4 M serotonin creatinine phosphate (5-hydroxy-
tryptamine). In addition, the following combinations of
serotonin and serotonin-specific chemicals were tested:
5.7 X 10~4 M serotonin plus 1.5 X 10~5 M cinanserin
(Squibb 10,643 cinnamanilide hydrochloride), a serotonin
antagonist; 5.7 X 10~4 M serotonin plus 1.3 X 10~3 M
fluoxetine, a serotonin uptake inhibitor; 10~4 g/rnl p-
chloroamphetamine (PCA), a serotonin releasing agent;
and 1 X 10~3 M 5-methoxytryptamine (5-MT), the pre-
cursor to serotonin (see Fuller, 1982). The control con-
sisted of filtered seawater alone. All solutions were pre-
pared in advance and frozen in glass vials in 50 ml aliquots
until time of use. For testing, vials were thawed and so-
lutions equilibrated to 10°C prior to filling the experi-
mental chamber. Intact specimens or isolated hepatopan-
creas tissue were then immersed in the test solution. Per-
meability of solutions to the site of action was not
considered to be a problem with this protocol because it
has been successfully used on euphausiids and shrimps
treated with serotonin, cinanserin, and other compounds
(Herring, 1976; Herring and Locket, 1978).
The kinetics of the luminescent responses were de-
scribed according to the following terms: latency, the time
period from presentation or termination of stimulus to
beginning of response: half rise, time from stimulus pre-
sentation to half maximum response amplitude; half decay
time, time from stimulus termination to half maximum
response amplitude. Unless otherwise stated, values are
stated as mean ± standard error of the mean.
Parametric statistical tests included the two-sample T
test and one-factor analysis of variance, while the Mann-
Whitney U test and Kruskal-Wallis test were used for
nonparametric comparisons. All statistical tests were per-
formed using Statview software (Abacus Concepts, Inc.).
Results
Photic stimulation
The luminescent response of restrained specimens of
Sergestes similis to photic stimulation depended upon
the degree of recent light exposure. Previously untested
animals responded differently from counterilluminating
specimens.
The typical counterillumination response to a dim
photic stimulus (Fig. 3A) displayed a latency of 2 s and
reached half maximum intensity within 13 s (Table I).
Generally, steady-state emission was achieved within
25 s. Luminescent intensity remained stable while the
stimulus was maintained. Upon termination of the stim-
ulus, luminescence was rapidly extinguished after a latency
of 1 s (Table I). The kinetics of the counterillumination
responses in the present study were similar to those pre-
viously measured (Warner el a/., 1979).
These responses were not present in previously un-
tested, dark acclimated specimens of 5. similis. There was
a latency period of several minutes during which no lu-
BIOIUMINESCENCE IN A SERGESTID SHRIMP
395
B
D
10 sec
Figure 3. Comparison of countenllummation and slow photic induction
of bioluminescence. For each trace of the chart recorder record, the
upper trace is the bioluminescence record, with an upward deflection
indicated increasing levels of light emission; the lower trace monitors
stimulus illumination, with a solid bar indicating no photic stimulus and
a clear bar representing stimulus "on." (A) Countenllummation in re-
sponse to stimulus illumination of 1 x 10~3 MW cm"2. Steady lumines-
cence was produced only when the light stimulus was present, and was
rapidly extinguished at the end of stimulation. (B-E) Slow photic in-
duction of bioluminescence in a previously untested, dark-acclimated
animal, illuminated by a maintained light stimulus with an intensity of
2 x 10~4 jiW cnT: (clear bar). Dim bioluminescence slowly increased
in intensity until termination of the stimulus midway through (E) (dark
bar), which produced a rapid extinguishing of luminescence. The sub-
sequent photic stimulus (clear bar) resulted in rapid "on" kinetics similar
to those of countenllummation in (A).
minescence was produced; subsequently, light emission
slowly increased (Fig. 3). Based on photomultiplier mea-
surements, light emission was first detected 3.3 ± 0.7 min
(range 2-5 min) after stimulus initiation. Luminescence
reached half maximum intensity after 12 min; maximum
steady light output occurred after approximately 25 min
of illumination (Table I).
Image intensification confirmed that light emission
originated from the organs of Pesta (Figs. IB, 4). Based
on observations of 31 previously untested animals, lu-
minescence was induced in the anterior organs 2.4 ± 0.3
min (mean ± standard error) after presentation of the
light stimulus, and in the posterior organs 3.6 ± 0.5 min
after the beginning of stimulation. Even though there was
no statistical significance to the earlier onset of emission
by the anterior organs (paired-sample / test, t = 0.86, P
> 0.20), this trend was observed in more than 60% of the
specimens tested. In most cases, the anterior light organs
were the brightest, and light emission from the midgastric
organs was very dim if detected at all.
Once luminescence was induced, an animal was capable
of subsequent responses with fast kinetics typical of the
counterillumination response. Termination of the initial
photic stimulus resulted in a rapid extinguishing of lu-
minescence (Fig. 3E) after a latency of 1 s. The kinetics
of the induction "off1 response did not significantly differ
from those of the counterillumination "off" response
(Table I). All subsequent photic stimulation resulted in
light emission with rapid response kinetics. An "on" re-
sponse latency of 2 s and time to maximum intensity of
25 s was similar to those of conventional counterillumi-
nation responses. Image intensifier observations under
these conditions indicated that once luminescence was
induced, the light organs invariably responded synchro-
nously to stimulus "on" and "off."
Photic stimulation was needed not only for induction
of the counterillumination response, but also to maintain
this state. Preliminary observations indicated that after
1 h of darkness, a previously counterilluminating animal
underwent a new induction process similar to those of
untested specimens. Subsequent to this, counterillumi-
nation was regained.
Eye glow, indicative of the dark-adapted eye state (Ball
et ai, 1986), was observed prior to testing in 5 of 6 dark-
maintained specimens, but was absent after testing. An-
imals adapted to a light intensity of 1 X 10~: /uW cm :
(an intensity higher than that present in their depth range;
Clarke. 1966) did not exhibit eye glow (0 of 4 specimens),
suggesting that the eye is light adapted at this level of
illumination. The threshold for light adaptation was not
determined.
Chemical stimulation
Serotonin was the only neurotransmitter tested that was
effective in producing bioluminescence (Fig. 5). Maxi-
mum levels of light emission from intact animals im-
mersed in 5.7 X 10"4 A/ serotonin were significantly dif-
396
M. I. LATZ AND J. F. CASE
Table I
kinetics ol the luminescent responses <>/ Sergestes similis
Condition
"On" latcncv
(s)*
Half rise time
(s)*
"O1T latency
(s)t
Half decay time
(S)f
Induction
Counterillumination
198.0 ± 39.0(4)
2.2 ± 0.2(11)
750.0 ± 86.6 (3)
12.8 ± 1.2(11)
1.3 ± 0.8 (2)
0.9 ± 0.1 (10)
2.5 ± 1.5(2)
1.6 ±0.1 (11)
* Mean values for induction and counterillunimation conditions are significantly different (Two-sample T test, t > 9.497, P < 0.001 ).
t No significant difference between means for test conditions (t < 1.584. P > 0. 10).
Values represent means with standard errors of the mean; number of observations given in parentheses.
ferent from seawater controls (Mann- Whitney U test, U
= 42, P < 0.01 ). The average temporal response consisted
of a latency of 6.0 ± 0.5 min followed by a slow increase
to maximum intensity that was reached in 26.7 ± 3.5 min
(Fig. 6B). These response kinetics are similar to those for
luminescent induction by photic stimulation.
Treatment with the neurotransmitters acetylcholine,
GABA, L-glutamic acid, and norepinephrine did not re-
sult in levels of bioluminescence significantly different
from seawater controls (Mann- Whitney U-test, P > 0. 1 )
(Fig. 5).
The specificity of serotonin in stimulating biolumines-
cence was further investigated (Fig. 5). A solution of se-
rotonin and tluoxetine, a serotonin uptake inhibitor, did
not produce significantly higher levels of light emission
compared to serotonin alone (Mann-Whitney U test. U
= 18, P > 0.5). nor did a solution of serotonin and cin-
anserin, a serotonin antagonist, produce significantly
lower levels of light emission (Mann-Whitney U test, U
= 18, P > 0. 1 ). There was no difference in the response
latencies for these conditions from that of serotonin alone
(Kruskal-Wallis test, P > 0.05). Treatment with PCA, a
serotonin releasing agent, and 10 3 M 5-MT, a serotonin
agonist, did not result in significant production of lumi-
nescence (Mann-Whitney U-test, P> 0.2).
Isolated hepatopancreas tissue containing the light or-
gans produced background levels of luminescence that
were significantly higher than for intact animal seawater
controls (Fig. 5, 6C) (Mann- Whitney U-test, U = 55, P
< 0.01). Treatment of the isolated tissue with serotonin
did not significantly alter the control glowing (Mann-
Whitney U-test, U = 25, P > 0.5).
Figure 4. Image intensifier views of slow photic induction of bioluminescence. Photographs of ventral
views of an animal were made from single fields of the video record. (A) View of cephalothorax of an intact
restrained specimen (clamp at bottom) under dim red light illumination. Anterior end is up. Bioluminescence
was observed (B) 2 min. (C) 3 min, and (D) 6.5 min after the beginning of maintained light stimulation,
showing emission initially from the anterior organs of Pesta, and then dimmer emission from the posterior
light organ. For (B-D) the stimulus light was briefly extinguished for photographic documentation. Scale
bar in (A) = 5 mm.
BIOLUMINESCENCE IN A SERGESTID SHRIMP
397
£
co
5
LU
tr
2400
2200
2000
1800
1600
1400
1200
800
600
400
200
8
.6 5
E £ E 2
S S B 2
2 S 2 1
E O
o y
E S
is £
^ 9-
Intact Isolated
Hepato-
pancreas
Figure 5. The effect of chemical treatment on bioluminescence. For
each experiment, the maximum relative intensity of light emission in
the initial 35 min of stimulation was determined. Concentrations of
solutions are given in Materials and Methods. Mean intensities with
standard errors of the mean are shown for each condition: the number
ol treatments is displayed abo\e each bar. An (*) indicates that the ex-
perimental treatment produced bioluminescence significantly different
from seawater control levels in intact animals (Mann-Whitne\ U test.
P < 0.05).
Effect ol 'eyestalk manipulation
Squeezing or ablating the eyestalks of previously un-
tested animals immediately evoked luminescence. It was
not possible to obtain response latency values as the PMT
shutter was closed during the eyestalk manipulation.
However, when the shutter was opened 5 s following the
procedure, light emission was present.
Serotonin treatment was effective in animals with
ablated eyestalks (Fig. 7). In one experiment, eyestalk-
less animals treated with serotonin produced a higher
intensity of light emission than the serotonin-stimulated
luminescence of intact specimens (Mann-Whitney U
test, U = 40, P < 0.01). Squeezing a single eyestalk of
intact serotonin-induced luminescing animals imme-
diately increased light emission by more than a factor
of2(n = 3).
Discussion
The responses to light of previously untested, dark-
maintained specimens of Sergestes ximilis clearly differed
from the typical counterillumination responses ascribed
to this species (Warner el til.. 1979). Previously untested
animals generated no detectable luminescence for several
minutes after initial photic stimulation; subsequently, light
emission increased to a maximum and steady level ap-
proximately 25 min later. However, once induced, sub-
sequent luminescent responses displayed the rapid kinetics
typical of the counterillumination response. The different
30 r A
0 5 10 15 20 25 30 35
TIME (min)
Figure 6. The comparison of slow induction of bioluminescence by
photic and serotonin stimulation. The intensity of emission is shown as
a function of time of stimulation. (A) The response of an uninduced
dark -adapted specimen to initial photic stimulation with a light intensity
of 4 X 10~4 /jW cm~:. Bioluminescence is expressed in relative units.
(B-C) Luminescent responses (expressed as PMT anode current in ^A)
to treatment with 5.7 x \Q~* M serotonin and seawater controls. (B)
Responses of intact animals. Serotonin was effective in producing a slow
rise in light emission (solid circles), while seawater was ineffective (open
circles). (C) Tests with isolated hepatopancreas tissue containing the lu-
minescent organs of Pesta. Serotonin treatment (solid circles) did not
increase luminescence above initial high background levels. Scawatci
control levels (open circles) were higher than controls for intact animals
(open circles in B).
398
M. I. LATZ AND J. F. CASE
11
8.8
UJ 6.6
LLJ
LLJ
rr
4.4
2.2
0
Serotonin Seawater
Intact
Serotonin Seawater
Eyestalkless
Figure 7. The effect of bilateral eyestalk ablation on bioluminescence.
The mean intensity of maximum light emission produced in the first 35
min of stimulation is shown along with standard errors of the mean. All
serotonin treatments were significantly different from seawater controls
( Mann-Whitney U test, P < 0.0 1 ). The serotonin response of eyestalkless
animals was significantly different from that of intact specimens (Mann-
Whitney U test, P < 0.0 1 ).
kinetics of the induction and counterillumination re-
sponses suggest dual control mechanisms regulating light
emission.
One mechanism for control of bioluminescence in 5.
similis may involve a neuronal pathway. Evidence for
neural control includes: ( 1 ) the immediate production of
luminescence upon eyestalk ablation of previously un-
tested animals, and (2) the immediate increase in light
emission following pinching of an eyestalk of an actively
luminescing specimen.
The kinetics of the counterillumination response in S.
.similis are consonant with those of neurally controlled
systems. Direct electrical stimulation of the spinal cord
of myctophid fishes results in luminescent response la-
tencies of 15 s or less (Anctil, 1972; Barnes and Case,
1974). Intact counterilluminating myctophids exhibit av-
erage response latencies of 1 to 18.5 s (Case et al.. 1977),
with a half rise time of 12 s and a half decay time of 1-2
s (Young el al., 1979). Even though at present there is no
morphological evidence for innervation of the organs of
Pesta of S. similis (Herring, 1981), the kinetics of the
counterillumination response of S. similis (half rise time
of 12.8 s, half decay time of 1.6 s) are similar to those of
the neurally controlled myctophid control system.
The long latency and slow increase in emission intensity
during the induction process suggest a different control
mechanism active during this period. Several features of
the induction process support the involvement of a blood-
born or neurosecretory pathway: ( 1 ) Photic induction of
bioluminescence occurred at a similar rate to chromato-
phore pigment dispersion in crustaceans, where an in-
crease in illumination causes release of erythrophore pig-
ment dispersing hormone (reviewed by Rao, 1985). (2)
Bioluminescence is stimulated by serotonin, which is a
known crustacean hormone releasing factor (reviewed by
Rao, 1985; Fingerman, 1987). (3) The loss of the coun-
terillumination state in S. simi/is after dark re-adaptation
may be due to the clearing of a blood-born substance,
similar to the return to the dark-adapted state of the crus-
tacean eye via gradual clearing of light-adapting hormone
from the hemolymph (Brown el al., 1952). Initial obser-
vations confirmed that eye glow in untested specimens of
5. similis, which indicated a dark-adapted eye state (Ball
el ill.. 1986). was absent after testing, indicating a change
to the light-adapted eye state. (4) The induction process
did not appear to involve the light-producing ability of
the photogenic cells, because the light organs of uninduced
specimens produced immediate luminescence upon
squeezing or ablating the eyestalks.
A bioluminescence induction process has not been de-
scribed for other counterilluminating midwater animals.
Some species of shallow-living leiognathid fishes of the
Indo-Pacific exhibit an initial slow rise in light emission,
proportional to the previous period of dark adaptation,
although this is due to chromatophore modulation of light
organ transparency rather than physiological regulation
of the production of luminescence (McFall-Ngai and
Morin. 1991; McFall-Ngai, pers. comm.). Perhaps a more
analogous phenomenon is arousal in the firefly Photitris,
which, if stimulated during daytime, requires 15 to 30 s
before flashes can be generated. During this period, the
light organ glows with increasing intensity and, finally,
flashing capability is established just after a rapid quench-
ing of the glow (Case and Buck. 1963).
The adaptive significance of an uninduced state and
the slow induction of bioluminescence is obscure. S. sim-
ilis does perform diurnal vertical migrations (e.g.. Clarke,
1966; Pearcy and Forss, 1969; Omori and Cluck, 1979)
during which it apparently follows a particular isolume
(Clarke, 1 966 ). Continuous exposure to dim downwelling
illumination would serve to maintain animals in the active
counterilluminating condition. On moonless nights, when
levels of downwelling illumination would be undetectable
and counterillumination unnecessary, animals would re-
vert to the uninduced condition. This might prevent in-
advertent luminescent responses to the luminescent dis-
plays of other animals and thereby reduce the chance of
being detected by predators. Although S. similis can re-
spond to light pulses as short as 2 s in duration (Warner
et al., 1979), it is not known if it responds to shorter du-
ration stimuli typical of luminescent flashes.
The role of light in the induction of counterillumination
in S. similis differs from light pulses that produce bursts
of luminescence in some organisms. For the shrimp Tluil-
ussoctiris (Herring and Barnes, 1976), copepod Methdia
longa (Lapota et al.. 1986). ostracods (Tsuji et al.. 1970:
BIOLUMINESCENCE IN A SERGESTID SHRIMP
399
Morin. 1986), and pyrosomes (Bowlby el al. 1990) a
photic stimulus acts as a trigger to release luminescent
behavior. In contrast, the long time course ofluminescent
induction in S. similis suggests a longer-term change in
physiological state occurring during the induction process.
Salient features of the S. similis luminescent system
are similar to those of euphausiids. In euphausiids, light
emission is stimulated by bright light or strobe illumi-
nation after a latency of several minutes. Serotonin is the
only neurotransmitter that stimulates light emission in
euphausiids, with a latency of 5 to 15 min (reviewed by
Herring and Locket, 1978). This response occurs only in
intact animals; isolated photophores treated with sero-
tonin do not luminesce (Herring and Locket, 1978). Al-
though the euphausiid control system has not been fully
elucidated, it is believed to involve control of blood flow
through the photophores by innervated sphincters
(Harvey, 1977; Herring and Locket, 1978).
Serotonin is present in the tissues of many marine in-
vertebrates (reviewed by Walker, 1984), and has been de-
tected in the eyestalks, cerebral ganglia (brain), ventral
nerve cord, and hemolymph of Crustacea (e.g., Fingerman
et al., 1974; Elofsson el al. 1982; Laxmyr, 1984). It is
well known to act on the crustacean neuromuscular junc-
tion by increasing neurotransmitter release (reviewed by
Kravitz et al.. 1985). Serotonin also acts on neurosecretory
cell terminals in the sinus gland of the crustacean eyestalk.
It mediates the release of a putative neurodepressing hor-
mone, a putative molt-inhibiting hormone, the hypergly-
cemic hormone, and a red chromatophore pigment dis-
persing hormone from neurosecretory cells in the eyestalk
(reviewed by Rao, 1985; Fingerman, 1987). The pigment
dispersing hormone is effective only in intact animals;
direct treatment of serotonin on erythrophores in isolated
legs or carapace has no effect (Nagabhushanam et al.,
1987; reviewed by Fingerman, 1987). This hormone also
acts to cause migration of the retinal distal pigment to the
light-adapted state (Kleinholz, 1975).
There are no marine luminescent systems in which di-
rect hormonal control of light emission has been dem-
onstrated. Direct innervation of squid and euphausiid light
organs occurs even when the photophores receive a rich
blood supply through an extensive capillary network (Ar-
nold and Young, 1974; Herring and Locket, 1978). Con-
trol of leiognathid bioluminescence through muscular
shutters may be fine-tuned through the action of chro-
matophores with slow response times (McFall-Ngai and
Morin. 1 99 1 ), which are presumably under neural control.
The present data suggest at least two sites involved in
the control of bioluminescence in S. similis. The eyestalk
contains the photoreceptors that detect downward-di-
rected illumination, and associated efferent neural or
neurosecretory cells. The responses of eyestalkless animals
to serotonin suggest an additional control site, possibly
located in the central ganglia. Furthermore, spontaneous
light emission from isolated light organs suggests inhibi-
tory control of light emission. The close coupling of vision
and bioluminescence in 5". similis may be achieved via a
hormonal component simultaneously active in the visual
and luminescent systems.
Acknowledgments
We are most grateful for the assistance of S. Willason,
K. Johnson, T. Frank, and M. Jess during nocturnal col-
lecting, M. Jess, D. Rupp, and G. Hallock with manuscript
preparation, and P. Herring, J. Morin, and M. Grober for
comments on the manuscript. Cinanserin was a gift of
E.R. Squibb and Sons, Inc. Supported by Office of Naval
Research contracts N00014-75-C-0242 (to JFC), N00014-
89-J-1477 (to MIL), and University of California faculty
research funds.
Literature Cited
Anctil, M. 1972. Stimulation of bioluminescence in lanternfishes
(Myctophidae). II. Can. J Zoo/. 50: 233-237.
Arnold, J. M., and R. E. Young. 1974. Ultrastructure of a cephalopod
photophore. I. structure of the photogenic tissue. Biol. Bull. 147:
507-521.
Ball, E. E., L. C. Kao, R. C. Stone, and M. F. Land. 1986. Eye structure
and optics in the pelagic shrimp Aceiessihogae (Decapods, Natantia.
Sergestidae) in relation to light-dark adaptation and natural history.
Phil. Trans. R Soc. Loud. B313: 251-270.
Barnes, A. T., and J. F. Case. 1974. The luminescence of lanternfish
(Myctophidae): spontaneous activity and responses to mechanical,
electrical, and chemical stimulation. J. Exp. Mar. Biol. Ecol. 15:
203-221.
Bowlby, M. R., E. A. Widder, and J. F. Case. 1990. Patterns of stim-
ulated bioluminescence in two pyrosomes (Tunicata: Pyrosomatidae).
Biol. Bull. 179: 340-350.
Brown, F. A., M. Fingerman, and M. N. Mines. 1952. Alterations in
the capacity for light and dark adaptation of the distal retinal pigment
of Palaemonetes Ptiyxiul. Zoo/. 25:230-239.
Case, J. F., and J. Buck. 1963. Control of flashing in fireflies. II. Role
of central nervous system. Biol. Bull. 125: 234-250.
Case, J. F., J. Warner, A. T. Barnes, and M. Lowenstine.
1977. Bioluminescence of lantern fish (Myctophidae) in response
to changes in light intensity. Nature 265: 179-181.
Clarke, W. D. 1963. Function of bioluminescence in mesopelagic or-
ganisms. Nature 198: 1244-1246.
Clarke, W. D. 1966. Bathyphotometric studies of the light regime of
organisms of the deep scattering layers. A.E.C. Research and Devel-
opment Rep. UC-48, Biology and Medicine. 47 pp.
Dennell, R. 1940. On the structure of the photophores of some decapod
Crustacea. Discovery Rep 20: 307-382.
Dilley, P. N., and P. J. Herring. 1974. The ocular light organ ofBalh-
othauma lyromma (Mollusca: Cephalopoda). J Zoo/. Loud 172:
81-100.
Elofsson, R., L. Laxmyr, K. Rosengren, and Ch. Hanson. 1982. Identi-
fication and quantitative measurements of biogenic amines and
DOPA in the central nervous system and haemolymph of the crayfish
Pacifastacus leniusculus (Crustacea). Conip. Biochem. Physiol. 71C:
195-201.
Fingerman, M. 1987. The endocrine mechanisms of crustaceans. J.
Crustacean Biol 7: 1-24.
400
M. I. LATZ AND J. F. CASE
Fingerman, M., \V. E. Julian, M. A. Spirtes, and R. M. Kostrezewa.
197-1. The presence of 5-hydro\ytryptamine in the eyestalks and
brain of the fiddler crab L'ca pngilalor, its quantitative modification
by pharmacological agents, and possible role as a neurotransmitter
in controlling the release of red pigment-dispersing hormone. Comp.
Gen. Pharmacol. 5: 299-303.
Fuller. K. \\ . 1982. Drugs acting on serotonergic neuronal systems.
Pp. 221-247 in Biology of Serotonergic Transmission. N. N. Osborne.
ed. John Wiley & Sons, New York.
Harvey, B. J. 1977. Circulation and dioptric apparatus in the photo-
phorcs of Euphausia pacijica: some ultrastructural observations. Can
J Zool. 55: 884-889.
Herring. P. J. 1976. Bioluminescence in decapod Crustacea. J Mar
Biol Assoc IK. 56: 1029-1047.
Herring. P. J. 1977. Luminescence in cephalopods and fish. Symp.
Zool. Sot:. Loud. 38: 127-159.
Herring, P. J. 1981. The comparative morphology of hepatic photo-
phores in decapod Crustacea. ./. Mar. Biol Assoc. U.K. 61: 723-737.
1 lerring, P. J. 1982. Aspects of the bioluminescence of fishes. Oceanogr
Mm Biol Aim Rev 20:415-470.
Herring, P. J. 1983. The spectral characteristics of luminous marine
organisms. Prot: R Sot: Loud. 220: 183-217.
Herring. P. J.. and A. 1 . Barnes. 1976. Light-stimulated biolumines-
cence of riialassoeuris cnnitu (Dana) (Dccapoda, Candea). Criis-
laccunaM: MI7-1 10.
Herring. P. J.. and N. A. Locket. 1978. The luminescence and pho-
tophores of euphausiid crustaceans. J. /<>«/.. Loud 186: 431-462.
Kleinholz, L. H. 1975. Purified hormones from the crustacean eyestalk
and their physiological specificity. Nature 258: 256-257.
krautz, K. A.. B. Beltz, S. Glusman, M. Coy, R. Harris-Warrick, M.
Johnston, M. Livingstone, I. Schvtartz, and K. King Siwicki.
1985. The well-modulated lobster. The roles of serotonin, octo-
pamine. and proctolin in the lobster nervous system. Pp. 339-360
in Model Neural Networks and Behavior. A. I. Selverston, ed. Plenum
Press. New York.
Lapota. D., .1. R. Losee, and M. L. Geiger. 1986. Bioluminescence
displays induced by pulsed light. Limnol. Oceanogi: 31: 887-889.
Latz. M. I., and J. F. Case. 1982. Light organ and eyestalk compen-
sation to body tilt in the luminescent midwater shrimp,
similis. ./ £.v/> Biol. 98: 83-104.
Laxmyr, I,. 198-1. Biogenic amines and dopa in the central nervous
system of decapod crustaceans. Comp. Biochem. Physiol. 77C: 139-
143.
McFall-Ngai, M., and J. G. Morin. 1991. Camouflage by disruptive
illumination in Leiognathids. a family of shallow-water, biolumi-
nescent fishes. / Exp. Biol. 156: 1 19-137.
Morin, J. G. 1986. "Firefleas" of the sea: luminescent signaling in
marine ostracode crustaceans. The Florida Entomol 69: 105-121.
Nagabhushan, R., R. Sarojini, and S. M. Kandharkar. 1987. Action
of biogenicamine on crustacean chromatophores. Analysis of response
of erythrophores in the prawn Caridina rajad/iari to serotonin. J.
Amm Morphol. Physiol 34:83-88.
Omori, M., and D. Gluek. 1979. Life history and vertical migration of
the pelagic shrimp Sergcstes sinuli.s olfthe Southern California coast.
Fish. Bull 77: 183-198.
Pearcy, VV. G., and C. A. Forss. 1969. The oceanic shrimp Sergestes
similis off the Oregon coast. Limnot. Oceanogr. 14: 755-765.
Rao, K. R. 1985. Pigmentary Effectors. Pp. 395-462 in The Biology
of Crustacea, I '<>/. V. Integument. Pigments and Hormonal Processes,
D. E. Bliss, and L. H. Mantel, eds. Academic Press. Orlando. FL.
Tsuji. F. I., R. V. Lynch, and V. Haneda. 1970. Studies on the biolu-
minescence of the marine ostracod crustacean Cypridina serrata.
Biol Bull. 139:386-401.
Walker, R. J. 1984. 5-Hydroxytryptamine in invertebrates. Comp.
Biochem Physiol. 79C: 231-235.
Warner, J. A., M. I. Latz, and J. F. Case. 1979. Cryptic biolumines-
cence in a midwater shrimp. Science 203: 1 109-1 1 10.
Waterman, T. H. 1961. Light sensitivity and vision. Pp. 1-64 in The
I'lmiologv i>/ Criistiicen. T. H. Waterman, ed. Academic Press, New
York.
\\idder, E. A., M. I. Latz, and J. F. Case. 1983. Marine biolumines-
cence spectra measured with an optical multichannel detection sys-
tem. Biol Bull 165: 791-810.
Young. R. FL 1983. Oceanic bioluminescence: an overview of general
functions. Bull. Mm: Sci. 33: 829-845.
Young, R. E., C. F. F. Roper, and J. F. Walters. 1979. Eyes and ex-
traocular photoreceptors in midwater cephalopods and fishes: their
roles in detecting downwellmg light for counterillumination. Mm
Biol 51: 371-380.
Reference: Biol Bull 182: 401-408. (June, 1992)
Evidence for Ammonia as a Natural Cue for
Recruitment of Oyster Larvae to Oyster Beds
in a Georgia Salt Marsh
W. K. FITT1 AND S. L. COON2
Department of Zoology, University of Georgia, Athens, Georgia 30602 and2 Department
of Microbiology and Center of Marine Biotechnology, University of Maryland.
College Park. Maryland 20742
Abstract. Competent veliger larvae of the oysters Cras-
sostrea virginica and C. gigas exhibited settlement be-
havior when exposed to ammonia (NH3). The threshold
for this response decreased with increasing larval age. The
response of veligers to adult-conditioned seawater was
correlated with the concentration of NH3 in the seawater.
Although the concentrations of NH3 found in marsh water
flowing over oyster beds on Sapelo Island. Georgia, were
never high enough to elicit settlement behavior from oys-
ter larvae, the concentrations found near the substrate
were sufficient to induce settlement behavior in older lar-
vae of C. virginicn. In addition, dilution occurs during
sampling in the field and may lead one to underestimate,
by a factor of 1.7 to 3.5, the actual concentration of NH3
associated with surfaces. In conclusion. NH3 may be an
important environmental cue triggering settlement be-
havior of larval oysters, which, along with other substrate
cues, leads to cementation and metamorphosis.
Introduction
As a prelude to attachment and metamorphosis, veliger
larvae of oysters exhibit a set of specific behavioral re-
sponses known as settlement behavior. This behavior is
induced by a variety of chemicals, including L-3,4-dihy-
droxyphenylalanine (L-DOPA) which is known to activate
larval endogenous dopaminergic neural pathways (Coon
ct u/.. 1985; Bonar et ai. 1990). Recent laboratory ex-
periments showed that ammonia (NH3) also induces set-
tlement behavior of oyster larvae that are competent to
undergo metamorphosis (see Coon el ill.. 1990a). but by
Received 2 December 1991; accepted 6 March 1992.
a different mechanism, possibly involving pH-induced
depolarization of nerve cells (Coon et a/., 1990a.b). A
similar mechanism is thought to be involved in the am-
monia and bacterial induction of settlement and meta-
morphosis of echinoid larvae (Gilmour, 1991; pers. com.).
Although most organisms release ammonia as a by-prod-
uct of metabolism, and the anaerobic muds characterizing
oyster habitats usually contain high concentrations of
ammonia, the potential importance of ammonia from
these sources for settlement of oysters is not known. The
presence of bacterial films on surfaces is often correlated
with oyster settlement and metamorphosis, suggesting that
one or more cues associated with either the bacteria or
their released metabolites is important for recruitment
(Galtsoff, 1964: Crisp. 1967; Weiner et ill., 1989: Bonar
et ul.. 1990). In addition, marine bacteria isolated and
grown in laboratory cultures release soluble cues that in-
duce settlement and metamorphosis of oyster larvae (Fitt
et ul.. 1989. 1990). A variety of analyses of supernatants
from cultures ofShewanella colwelliana, a bacterial species
known to enhance oyster recruitment, showed that their
ability to induce settlement behavior of oyster larvae was
correlated with the concentration of NH3. not with that
of melanin nor of any other catechol-related intermediates
(Coon and Fitt. unpub.).
Oysters are gregarious, and although many attempts
have been made to elucidate factors responsible for the
settlement and metamorphosis of larvae around and on
adults, the responsible chemical cues have not yet been
conclusively identified (Cole and Knight-Jones. 1949;
Knight-Jones. 1952: Crisp. 1967;Hidu, 1969; Veitch and
Hidu. 1 971: Keck f/rt/.. 1971: Hidu el ai. 1978). Because
oysters and oyster reefs release ammonia (Mann. 1979:
401
402
W. K. FITT AND S. L. COON
Boucher and Boucher-Rodoni. 1985, 1988; Dame et ai.
1985, 1989), such release by congeners may trigger the
initial 'search behavior' portion of settlement and meta-
morphosis (see Coon et ai, 1990a, for model). Oyster
larvae exposed to adult extrapallial fluid exhibit "setting
behavior" within 10 min of exposure, the rapid response
characteristic of larval responses to NH3, but not to cat-
echolamines (Hidu et ai, 1978, Coon et ai, 1990b). In
addition, larvae induced with adult fluids have a higher
percentage attachment on shells than larvae exposed only
to seawater (Hidu et ai, 1978). Seawater that has been
conditioned by adult oysters also significantly increases
setting rates on cultch (in Hidu et ai, 1978). Whether
adult and juvenile oysters can produce enough NH3 to
induce settlement behavior, and whether larvae respond
to adult-produced NH3, has not previously been deter-
mined.
Another important but unanswered question is whether
enough ammonia is present on oyster reefs to affect larval
settlement behavior. The highly productive salt marshes
of the east coast of the United States are characterized by
organically rich mud containing partially decomposed
plant material. Oyster larvae are recruited to established
oyster reefs in these biologically rich and complex envi-
ronments. High concentrations of ammonia have been
measured from salt-marsh oyster habitats (Stevens, 1983;
Boucher and Boucher-Rodoni, 1985; Dame et ai, 1985.
1989), but these levels are typically 5-100 times lower
than those needed for induction of settlement behavior
by newly competent larvae (cf. Coon el ai, 1990b). How-
ever, oyster larvae (Fitt el ai, 1989; Coon el ai. 1990a),
as well as other invertebrate larvae (Knight-Jones, 1953;
Bayne, 1965; RittschoffV/ ai, 1984, 1986; Fitt and Hof-
mann, 1985; Crisp, 1988; Hadfield. 1977), become more
sensitive to morphogens as they age. For instance, the
threshold concentration of L-DOPA to which larvae of
the oyster Crassostrea gigas will respond decreases from
1(T5 M in early competency to 10~6 M three to four weeks
after the onset of competence (Coon and Fitt, unpub.).
In addition, during settlement, competent veligers inev-
itably encounter crevices and boundary layers on and near
surfaces where concentrations of chemicals originating
from these substrates are higher than those in the sur-
rounding seawater. No one, to our knowledge, has at-
tempted to look for chemical inducers of settlement of
oyster larvae in micro-habitats on oyster reefs.
We therefore hypothesized that oysters and oyster reefs
may produce iiigh enough concentrations of NH3 to trig-
ger settlement behavior of oyster larvae. We have tested
this by quantifying NH3 levels in and around oyster reefs
and comparing these with the responses of larvae of Cras-
sostrea virginica and C. gigas to NH3. In addition, we
investigated the decline, with larval age, in the threshold
concentration of NH3 required to induce settlement be-
havior. We report here that the concentration of NH3
measured in the extensive oyster bed habitats on Sapelo
Island, Georgia, overlaps the response range of competent
veliger larvae, suggesting that ammonia may be an inducer
of settlement behavior in nature.
Materials and Methods
Laboratory experiments
Larvae of the Pacific oyster Crassostrea gigas (Thun-
berg) were obtained from the Coast Oyster Company of
Quilcene, Washington, and those of the American oyster
C. virginica from either Horn Point Environmental Lab-
oratory, University of Maryland. St. George Oyster Com-
pany, Piney Point, Maryland, or Virginia Institute of Ma-
rine Science, Gloucester Point. Virginia. The larvae were
maintained in the laboratory as detailed in Coon et ai
(1990a).
Settlement behavior in veliger larvae of oysters includes
a well-characterized series of stereotyped maneuvers
(Coon et ai. 1990a). These include swimming with the
foot extended forward, followed by crawling on the sub-
strate in a progression of increasingly localized behaviors,
including reduced crawling speed and increased frequency
of turns. This behavior is initiated in competent veliger
larvae upon exposure to an appropriate soluble chemical
cue. and after perception of additional substrate cues may
result in cementation to the substrate. Settlement behavior
of competent larvae was monitored as previously de-
scribed (Coon et ai, 1990a). Between 20 and 50 larvae
were assayed in each well of 24-well tissue culture plates
(Falcon #3047) in a final volume of 1.0 ml at 20°C. Typ-
ically, six replicate wells were monitored for each exper-
imental condition. Settlement behavior of larvae actively
extending their foot was monitored in each well using a
dissecting microscope every 5 min. for a 1-min interval
over a 30-60 min period. Responses to concentrations of
NH4C1 (pH 7.8-8.0) ranging from 300 n.Mlo 9 ITL\/ were
determined with 19-30-day-old competent larvae of C.
virginica. Competent veliger larvae are defined as being
able to respond to external stimuli to trigger settlement
behavior and metamorphosis. This typically develops in
veliger larvae between 14-21 days post-fertilization and
is usually characterized by a well-developed foot and black
eye-spots. All veliger larvae used in experiments possessed
well-developed eye spots (Coon et ai, 1990a).
Some experiments were designed to monitor the set-
tlement behavior of competent veligers in response to
adult-conditioned seawater. Adult oysters (4-10 cm in
length) were scrubbed 1-2 days before the experiment
with a toothbrush and 10% hypochlorite solution to re-
move algae, invertebrates, and bacterial films. Cleaned
oysters were rinsed repeatedly to remove hypochlorite.
and were allowed to sit in fresh seawater for 24-48 h.
NATURAL SETTLEMENT CUE FOR OYSTERS
403
They were then placed in acid-washed glass petri dishes
containing 100 ml of 0.45 jum Millipore-filtered Instant
Ocean to begin the experiment. After 12-48 h, adult-con-
ditioned seawater was removed from the petri-dish and
assayed for NH3 concentration and its ability to induce
larval settlement behavior.
I'tckl experiment^
Field work was conducted in oyster beds (=reefs) in
South End Creek, adjacent to the Marine Institute of the
University of Georgia on the southern end of Sapelo Is-
land, Georgia. These oyster reefs, like many on the Geor-
gia coast, are characterized by high densities of oysters
that line the creek banks and bottoms in intertidal portions
of the marsh. Settlement of larvae onto these reefs occurs
sporadically throughout the spring, summer, and early
fall. This particular tidal creek drains a diked marsh, and
contains numerous oyster reefs along its 0.75-mile course
connecting it to Doboy Sound. Water samples were taken
during late spring (May 1990), summer (June 1990 and
1991. August 1991), and early fall (October 1989).
The characteristics of the marsh water overlying the
oyster reefs were determined from 5-101 water samples
collected hourly for 26 h during the diurnal tidal cycle
19-20 May 1990. Temperature and pH were measured
simultaneously with a portable temperature-compensated
pH meter immediately after collection of the water. A
refractometer was used to determine salinity, and oxygen
was measured with a calibrated YSI oxygen electrode
within 5 min of collection. Duplicate subsamples (5 ml)
were taken for ammonia determination (below). Tide
height at each collection time was calculated using a
marked rope weighted at one end. calibrated to the lowest
and highest water level.
Water samples for determination of natural levels of
ammonia associated with the oyster reefs were collected
with adjustable pipettors from three general habitats on
the oyster reef. First, the interface between creekwater
and the oyster reef was sampled at both high and low tides
from a canoe. Second, water was collected from tidepools
surrounded by oyster reefs. Third, small bodies of water
surrounding exposed oysters were sampled. In each hab-
itat, water was collected with the pipettors during low tide
from three sources: ( 1 ) horizontal and vertical surfaces of
adult and juvenile shells, with the resulting data combined
into a category called 'shell surface': (2) crevices between
oyster shells; and (3) open water near, or above, live oyster
reefs. In addition, some water samples were taken from
crevices and surfaces of submerged adult oysters during
an incoming tide (5-10 cm below the surface). Samples,
either 250 or 100 ^1, were diluted with deionized distilled
water before being assayed for total ammonia (NH3-NH4+)
as detailed below. Because the sampling procedures dis-
turb natural nutrient gradients, replicate samples were
taken from a similar habitat {i.e., shell surface, crevice
between shells) at the same sampling location, but not at
exactly the same position. In all cases, care was taken not
to disturb the water around oysters before collection.
Collecting water samples from shell surfaces and crev-
ices inevitably involves dilution of the immediate surface
water by the adjacent seawater. To estimate this dilution
effect during sampling, duplicate water samples of vol-
umes between 25 and 1000 ^' were collected from the
same surfaces. Ammonia was analyzed as described below
and plotted against volume sampled. A theoretical
boundary layer concentration was calculated by extrap-
olating the data back to a zero volume sample (by linear
regression). A dilution factor was calculated by dividing
the total NH3-NH4 + concentration at the theoretical zero
ml volume by the concentration sampled.
Ammonia determination
Total NHrNH4+ concentration was measured by a
modification (Wilkerson and Trench, 1986) of the phenol-
hypochlorite method (Liddicoat etai, 1975). Absorbance
of replicate samples were read on a spectrophotometer at
640 nm after a minimum 1 h incubation in the dark to
allow color development. Ammonia (NH3) concentration
was calculated from standard tables relating pH. salinity,
and temperature to the proportion of NH3 from the total
NH3-NH4+ content (Bower and Bidwell, 1978).
Results
Larval response to ammonia
Older larvae of C. virginica responded to lower con-
centrations of NH3 than newly competent larvae (Fig. 1 ).
The dose-response curves showed the typical peak in
maximum number of larvae exhibiting settlement behav-
ior in less than 10 min at the highest concentrations tested
(cf. Coon et a/., 1990b), extending to 20 min or longer as
the concentration of NH3 decreased (Fig. 1). The lowest
concentration of NH3 eliciting larval settlement behavior
(8.2 ± 3.3%. mean ± S.D.. of larvae responding, n = 6)
that was higher than controls (0%. n = 6) was 7.1 ju.U
(Fig. 2).
Larval response to adult-conditioned water
Competent larvae of both C. gigas and C. virginica
exhibited settlement behavior when exposed to adult-
conditioned seawater (Figs. 2. 3). The level of larval re-
sponse was similar to that expected from the amount of
NH3 found in the adult-conditioned seawater (Fig. 2). In
addition, the larval response to adult-conditioned water
increased in a predictable fashion when the NH3 concen-
tration in oyster-conditioned water was increased by rais-
404
W. K. FITT AND S. L. COON
19 days old
>
CD
Q)
D)
C
X
0)
d>
co
TO
c
Q>
O
OJ
Q.
0 10 20
Time (min)
Figure I . Percent of Crassostrea virginicu exhibiting settlement behavior
when exposed to various concentrations of NH4Cl as a function of ex-
posure time. All experiments were conducted at pH 7.8-8.0. Concen-
trations of NH, were approximately: solid squares = 180 nM NH,;
diamonds = 169 \j.M NH,; open squares = 1 12 nAl NH3; open circles
= 48 n.\t NH,: open triangles = 42 ^/NH3; closed triangles = 26 pAl
NH,; closed circles = 16 pM NH,. Data are means ± S.D., n = 6.
ing the pH of the conditioned water from 7.4 to 8.0, and
decreased by lowering the pH from 7.4 to 7.1 (Fig. 3).
of total NH,-NH4+ were measured during low tide, when
oxygen levels and pH were lowest. The total NH
concentrations in these samples did not exceed 20
and ammonia (NH3) was less than 1 n\l. The highest
NH3 concentrations in creek water were on an incoming
tide during the daytime (Fig. 4B). There were two low
points in NH3 concentration: ( 1 ) at peak high tide, when
total NH3-NH4+ was at its lowest in the seawater entering
the creek from Doboy Sound; and (2) on the outgoing
tide, when pH decreased relatively faster than the total
NHrNH4+ concentration increased (Fig. 4B).
When water samples were taken next to oyster shells,
in moving water on an incoming tide, NH3 concentrations
were similar to that of the creek water overlying the oyster
bed (Table I: 1, 2A3). However, water sampled from sur-
faces and between shells at low tide when the flow was
minimal often had higher levels of NH3 and total NHr
NH4+ than the overlying creek water (Table 1:1). Values
co
.c
0)
C
Q>
E
o>
0)
10
D>
C
X
CD
CD
co
en
co
CD
Q.
x
CO
40 n
30 -
20 -
10 -
1 0 0
200
300
400
1 00
200
300
400
Ammonia levels in an oyster heel
Total NHrNH4+ concentration in surface (0-20 cm)
creek water over an oyster bed on Sapelo Island in May
1990, varied with lide height and time of day (Fig. 4).
Total NH,-NH/ levels were inversely correlated with tide
height, pH. and oxygen (Fig. 4). Highest concentrations
Ammonia concentration
Figure 2. Maximum percent of larvae exhibiting settlement behavior
in response to either NH4C1 (filled circles) or adult-conditioned water
(open circles). (A) Crassostrea virginica. (30 days old), (B) C ,i,'W«. newly
competent larvae (19 days old). Data are means ± S.D. (n = 6) of the
maximum response, seen between 0 and 20 min. depending on the con-
centration of NHjCI.
NATURAL SETTLEMENT CUE FOR OYSTERS
405
pH.8.0 645(iMNH3
pH,7.4 183jiMNH3
7.1 B.S^MNHS
d>
D.
1 0
2 0
Time (min)
Figure 3. Percent ofCrassoslrea virginica exhibiting settlement behavior
in response to adult-conditioned seawater as a function of exposure time.
pH was adjusted before the experiment began in order to manipulate
the NH, concentration, as noted in the text. Data are means ± S.D.
(n = 6).
in Table I: I are from August 1 99 1; values from June 1 99 1
were similar to those found in August 1 99 1 . Samples taken
at earlier dates were lower, probably due to dilution from
surrounding water resulting from the larger volumes sam-
pled (> I. Oml), or due to seasonal differences. The highest
concentrations of total NH,-NH4+ were recorded between
oyster shells (crevices), and include numerous samples
exceeding 300 nM and a maximum concentration of 422
n.M total NH,-NH4+. The highest average ammonia (NH_,)
values were recorded on 2 August 1 99 1 around oysters
exposed on an incoming tide (mean = 6.6 ± 3.2 S.D. nM
NH3. range 3.9 to 1 1. 2 n\l, n = 9). Six out of 16 samples
taken from shell crevices on this date had concentrations
of NH, greater than 7.1 pM. Although the total NH,-
NH4+ concentrations were higher on some samples on
the previous afternoon. NH, concentration was always
lower than 7.1 ^.\f. because the pH of the outgoing tide
was so low.
Dilution factor
As smaller volumes were sampled from a surface of an
oyster shell, the measured total NH,-NH/ concentration
increased (Fig. 5). The calculated dilution factor varied
with sample volume, and was about 1.5 for volumes
> 250 n\ using a 1000 >A pipettor (Fig. 5). and 3.5 for 100
^1 samples using a 100 n\ pipettor (data not shown). These
dilution factors will obviously vary with type of habitat
and substrate sampled. Control samples from > 1 cm away
from a shell surface showed no significant difference in
ammonia concentration with sample volume. The highest
concentrations of NH, measured in the oyster beds (Table
I) are well within the minimum range needed for induc-
tion of settlement behavior of older larvae of C. virginica
(Fig. 2, above). Environmental concentrations of NH, as-
sociated with some shell surfaces, calculated using these
dilution factors, exceeded 30 pM, far surpassing the min-
imum values needed to elicit larval settlement behavior.
x
D.
Q.
Q.
>.
c
CO
.E
D)
9.0 -
8.8-
8.6-
8.4-
8.2-
8.0-
7.8-
7.6-
7.4
20 n
15 -
10 -
30 i
20 -
10 -
3 1
2 -
Oxygen
Tide heiqht
Temperature
25
20
-15
-10
-5
0
0.5
0.4
0.3
-0.2
-0.1
0.0
10
-8
-6
-4
-2
-30
-29
-28
-27
-26
-25
24
X
z
0.
a.
c
Q)
D)
>,
X
O
o
<U
Q.
E
<u
0600 1100 1600 2100 0200 0700
Time of day (h)
Figure 4. Characteristics of creek water from Sapelo Island on 19-20
May 1990 in relation to time of day. (A) Total NH,-NH4* and pH. (B)
NH, and total NH3-NH4*. (C) Salinity and oxygen. (D) Tide height and
temperature.
406
W K. FITT AND S. L. COON
Table 1
Concentration of total NH,-NHj [jiA/. means ± S.D (n)]. and the range of corersponding concentrations ofNHj
in oyster bed habitats nn Sapcln Island
Ambient seawater
Shell surface
Shell crevice
/ / August 1W1 (250 M/ samples)
A. Morning low tide (incoming tide), pH = 7.4-7.7
1. Creek
43.2+ 2.6(7)
52.1 ± 5.3(14)
70.5 ±32.5 (11)
Range (NH,):
0.8-0.9
0.9-1.2
0.9-3.2
2. Pool 1
101.0 ± 1.7(3)
111.3 ± 12.9(3)
165.4 ±66.3 (8)
Range (NH,)
2.0-2.0
2.0-2.5
2.0-5.4
B. Afternoon low tide (outgoing tide), pH = 7.3-7.4
1. Pool 1
81.9 ± 7.8(4)
76.5+ 3.5(2)
142.6 ± 56.8(9)
Range (NH,):
0.7-0.9
0.7-0.8
0.9-2.6
2. Pool 2
199.4 ± 14.0(2)
203.0 + 11.3 (2)
270.7 ± 82.7 (6)
Range (NH,):
1.9-2.1
2.0-2.1
2.1-4.2
3. Exposed oysters
n.d.
n.d.
256.6 + 55.5 (5)
Range (NH,):
n.d.
n.d.
1.7-3.3
//. .' August 1991 (100 nl samples)
A. Morning low tide (incoming tide), pH = 7.8-8.0
1. Pool 1
72.8(1)
n.d.
101. 3 ±38.4 (7)
Range (NH,):
2.8
n.d.
3.2-7.2*
2. Exposed oysters
n.d.
n.d.
169.9 ± 83.0(9)
Range (NH,):
n.d.
n.d.
3.9-11.2*
3. Between tides (reef underwater)
20.9 ± 6.7(5)
18.2 ± 1.1(2)
21.7 ± 2.1 (9)
Range (NH3):
0.5-1.1
0.7-0.7
0.7-1.0
* Six out of 16 water samples taken from these habitats had a high enough concentration of NH3 (>7.1
veliger larvae.
n.d. = no data.
to induce settlement behavior of
Discussion
Our goal in this study was to determine whether NH,
levels in nature are high enough to induce settlement be-
havior of veliger larvae of oysters. The data show that
concentrations of NH, close to oyster shells in oyster beds
at Sapelo Island reach concentrations at least as high as
the minimum concentration of NH, needed to induce
settlement behavior of larvae of C. virginica. In addition,
adult oysters produced enough NH, in laboratory exper-
iments to induce settlement behavior. These results sug-
gest that NH, concentrations in or near boundary layers
of surfaces in oyster beds may be at least partially re-
sponsible for triggering settlement behavior in nature.
The highest values for total NH,-NH4+ were found
during afternoon low tides in the summer, when temper-
atures are typically highest in the marsh (Table I). How-
ever, the interaction of pH of the ambient seawater and
total NH3-NH4+ (Fig. 4, Table I) combined to give con-
centrations of NH, that were higher on incoming than
outgoing tides. If competent oyster larvae are present in
the water column, and if NH, is one of the cues to which
they respond in nature as suggested in this study, then
one might expect oyster larvae to settle during incoming
tides rather than on outgoing tides. An alternative scenario
might find competent larvae in areas of quiescent water
on oyster reefs during low tide, where levels of NH, may
become very high. There are currently no convincing data
indicating that part of the tidal cycle, or time of day, when
oyster larvae tend to set.
Other chemical cues also trigger settlement behavior of
oyster larvae. A number of catecholamines, including L-
DOPA and norepinephrine, induce classic veliger settle-
ment behavior and subsequent metamorphosis (Coon el
ai, 1985, 1990a). Treatment with methylamine and other
weak bases also induces settlement behavior (Coon et a/.,
1990b). None of these compounds has been found in oys-
ter beds, but there is evidence from experiments in salt
marshes that other soluble cues may exist and play a role
in oyster settlement. Zimmer-Faust ( 1 990) and Tamburri
( 1990) found differences in larval swimming behavior in
marsh water containing sub-threshold concentrations of
NH3. The relationship between these changes in swim-
ming behavior and the specific behaviors involved in set-
tlement (e.g.. foot extension, crawling, and turning) are
unclear. While these other soluble cues may be important
in oyster recruitment, their identity and characteristics
are virtually unstudied.
Chemical induction of settlement behavior modifies
veliger movement in such a way as to bring competent
larvae into physical contact with potential substrates for
NATURAL SETTLEMENT CUE FOR OYSTERS
407
110 -
c 100-
90 -
I
Z
30-
70 -
60
200
400
600
1000
Volume sampled (ul)
Figure 5. Relation of total NH,-NH4+ measured to volume of water
sampled from oyster-shell surfaces on Sapelo Island using a 1 000 ^l
pipettor.
attachment and metamorphosis. Researchers have spec-
ulated for years about the characteristics of substrates
suitable for oyster attachment and metamorphosis, but
only the presence of other larvae or adults (Cole and
Knight-Jones, 1 949; Knight Jones, 1952; Crisp, 1974) and
biofilms (Crisp and Ryland, 1 960; Galtsoff, 1 964; Weiner
etal.. 1989) have convincingly correlated with higher oys-
ter set. The molecular factors associated with conspecifics
in nature are not known, but as shown here may involve
the production of ammonia. Our data show that larval
behavior can be altered by changing the availability of
NH3 in adult-conditioned water by changing pH. Others
have found a settlement-behavior inductive factor in
adult-conditioned water, but have been unable to identify
it (references in Hidu el at.. 1978).
Settlement of oyster larvae involves two basic steps. ( 1 )
Settlement behavior triggered by soluble cues that act to
bring the larvae in contact with surfaces, and (2) cemen-
tation and subsequent metamorphosis triggered by un-
known cues associated with surfaces (Coon el ai, 1990a).
The latter appear to be related to biofilms, but few ex-
periments have addressed this relationship (cf. Walch el
ai. 1987; Labare and Weiner, 1990). Many of the early
results showing more set on cultch coated with oyster ex-
tracts than control cultch (references in introduction) may
have been due to higher numbers of resulting bacteria,
and thus higher concentrations of NH3 as well.
Ammonia-induced settlement behavior does not by it-
self result in subsequent attachment and metamorphosis
in laboratory experiments (Coon el ai. 1990b; unpub.).
Experiments demonstrating this were performed in plastic
cell-culture plates, previously shown to be sub-optimal
setting surfaces for oyster larvae (Coon et a!., 1990a). Be-
cause larvae that are induced with NH, characteristically
habituate to that stimulus in less than 30 min and then
resume normal swimming behavior (Coon et al., 1990b).
we hypothesize that, in the laboratory, they do not spend
enough time in contact with this substrate to induce set-
tlement. Such a phenomenon may also occur in com-
petent larvae in nature, where the importance of selecting
among a variety of settlement sites may be crucial to sur-
vival. Such a process may be part of the mechanism by
which veliger larvae settle on premium substrates, such
as congener shells, more frequently than suboptimal sub-
strates, such as mud. The substrate factors important in
triggering final attachment and metamorphosis are not
currently known.
Acknowledgments
The authors thank the following oyster hatcheries for
veliger larvae: Coast Oyster Company of Quilcene, Wash-
ington; Horn Point Environmental Laboratory, Univer-
sity of Maryland; St. George Oyster Company, Piney
Point, Maryland; and Virginia Institute of Marine Science,
Gloucester Point, Virginia. We thank Doug Haymans,
Dr. Abdelmonem Khalil, and students in the 1990 Eco-
logical Physiology course at the University of Georgia for
help in collecting some of the data presented in this paper
and thank Dr. Dick Zimmer-Faust for information on
his current research. We also thank the staff at the Uni-
versity of Georgia Marine Institute on Sapelo Island for
logistical support of the field work. Portions of this work
were supported by the Sea Grant Program of Georgia,
and the National Science Foundation (DCB-9 108074 to
WKF). Contribution #191, Center of Marine Biotech-
nology, University of Maryland.
Literature Cited
Bayne, B. L. 1965. Growth and the delay of metamorphosis of the
larvae of Mytilus edulis (L.) Ophelia 2: 1-47.
Bonar, D. B., S. L. Coon, M. Walch, R. M. Weiner, and W. K. Fitt.
1990. Control of oyster settlement and metamorphosis by endog-
enous and exogenous chemical cues. Bull. Mar. Sci 46: 484-498.
Boucher, G., and R. Boucher-Rodoni. 1985. Fluctuations des nutriments
au cours de la maree sur les pares ostreicoles de la nviere Penze
(Nord-Fimstere). Hydrobiologia 123: 251-261.
Boucher, G., and R. Boucher-Rodoni. 1988. In silu measurement of
respiratory metabolism and nitrogen fluxes at the interface of oyster
beds. Mar. Ecol. Prog. Ser 44: 229-238.
Bower, C. E., and J. P. Bidwell. 1978. lonization of ammonia in sea-
water: effects of temperature, pH, and salinity. / Fish. Res. Board
Can. 35: 1012-1016.
Cole, H. A., and E. W. Knight-Jones. 1949. The setting behaviour of
larvae of the European flat oyster. Ostrea edulis L. and its influence
on methods of cultivation and spat collection. Fishery Invest.. Land.
Ser. II. 17: 1-39.
408
W. K. FITT AND S. L. COON
Coon, S. L., D. B. Bonar, and R. M. Weiner. 1985. Induction of set-
tlement and metamorphosis of the Pacific oyster, Crassosirea gigas
(Thunberg), by L-DOPA and catecholamines. J. Exp. Mar. Biol.
Ecol. 94:211-221.
Coon, S. L., W. K. Fitt, and D. B. Bonar. I990a. Competence and
delay of metamorphosis in the Pacific oyster, Crassosirea gigas. Mar.
Biol 106: 3~9-3S7.
Coon, S. I... M. Walch, \V. K. Fitt, R. M. Weiner, and D. B. Bonar.
199(lb. Ammonia induces settlement behavior in oyster larvae. Biol.
Bull 179:297-303.
Crisp, D. J. 1967. Chemical factors inducing settlement in Crassosirea
virginica (Gmelin). J. Anini. Ecol. 36: 329-335.
Crisp, D. J. 1974. Factors influencing the settlement of marine inver-
tebrate larvae. Pp. 1 77-265 in Chemoreception in Marine Organisms.
P. T. Grant and A. M. Mackie. eds.. Academic Press, London.
Crisp, D. J. 1988. Reduced discrimination of laboratory-reared cyprids
of the barnacle Balanus amplulnlc amphilriie Darwin, Crustacea
Cirripedia, with a description of a common abnormality. Pp. 409-
432 in Marine Biodeterioration, M. F. Thompson. R. Sarojini, and
R. Nagabushanam, eds.. Oxford and IBH Publ. Co.. New Delhi.
Crisp, D. J., and J.S. Ryland, 1960. Influence of filming and of surface
texture on the settlement of marine organisms, \atitre 185: 1 19.
Dame, R. F., T. G. Wolaver, and S. M. Libes. 1985. The summer
uptake and release of nitrogen by an interstitial oyster reef. Nctli. J
Sea Res. 19: 265-268.
Dame, R. F., J. D. Spurrier, and T. G. Wolaver. 1989. Carbon, nitrogen
and phosphorus processing by an oyster reef Mar Ecol. 54: 249-
256.
Fitt, W. K., and D. K. Hofmann. 1985. Chemical induction of settlement
and metamorphosis of planulae and buds of the reef-dwelling coel-
enterate Cassiopeia andromeda. Proc. 5th Intl. Coral ReefCongr. 5:
239-244.
Fitt, W. K., M. P. Labare, W. C. Fuqua, M. Walch, S. L. Coon, D. B.
Bonar, R. R. Colwell, and R. M. Weiner. 1989. Factors influencing
bacterial production of inducers ot settlement behavior of larvae of
the oyster Crassosirea gigas. Microb. Ecol. 17: 287-298.
Fitt, W. K., S. L. Coon, M. Walch, R. M. Weiner, R. R. Colwell, and
D. B. Bonar. 1990. Settlement behavior and metamorphosis of
oyster larvae of Crassostrea gigas in response to bacterial supernatants.
Mar Biol 106: 389-394.
Galtsoff, P. S. 1964. The American oyster, Crassosirea virginica
Gmelin. /-Wi. Bull Fish \ViIdlileServ. U. S 64: 1-480.
Gilmour, T. H. J. 1991. Induction of metamorphosis of echmoid larvae.
Am Zool 31: 105A.
lladfield, M. G. 1977. Chemical interactions in larval settling of a
marine gastropod. Pp. 403-413 in Marine Natural Products Chem-
istry. D. J. Faulkner and W. H. Fenical, eds.. Plenum Publ. Corp.,
NY.
1 lidu. II. 1969. Gregarious setting in the American oyster Crassostrea
virgmica Gmelin. Chesapeake Set 10: 85-92.
I lidu, H., W. G. Valleau, and F. P. Veitch. 1978. Gregarious setting
in European and American oysters — response to surface chemistry
i'.?. waterborne pheromones. Proc. Nail. Shellfisheries Assoc. 68: 11-
16.
Keck, R., D. Maurer, J. D. Kauer, and W. A. Sheppard. 1971. Chemical
stimulants affecting larval settlement in the American oyster. Proc.
Nail Shellfisheries Assoc. 61: 24-28.
Knight-Jones, E. W. 1952. Reproduction of oysters in the Rivers
Crouch and Roach, Essex, during 1947, 1948, and 1949. Fishery
Invest . Loud. Ser 11 18: 1-48.
Knight-Jones, E. W. 1953. Decreased discrimination during setting
after prolonged planktonic life in larvae of Sptrorbis borealis (Ser-
pulidae). J. Mar. Biol Assoc. U.K.. 32: 337-345.
Labare, M. P., and R. M. Weiner. 1990. Interactions between She-
\\anella colwelliana. oyster larvae, and hydrophobic organophosphate
pesticides. Applied Environ Microhm/. 56: 3817-3821.
Liddicoat, M. I., I. S. Tibbits, and E. Butler. 1975. The determination
of ammonia in seawater. Limnol. Oceanogr. 20: 131-132.
Mann, R. 1979. Some biochemical and physiological aspects of growth
and gametogenesis in Crassostrea gigas and Ostrea edulis grown at
sustained elevated temperatures. ./. Mar. Biol. Assoc. U.K. 59: 95-
110.
Rittschof, D., E. S. Branscomb, and J. D. Costlow. 1984. Settlement
and behavior in relation to flow and surface in larval barnacles. Bal-
anns amphilriie Darwin. J. Exp. Mar. Biol. Ecol. 82: 131-146.
Rittschof, D., J. Maki, R. Mitchell, and J. D. Costlow. 1986. Ion and
neuropharmacological studies of barnacle settlement. Nelh. J. Sea
Res. 20: 269-275.
Stevens, S. A. 1983. Ecology of Intertidal Oyster Reefs: Food. Distri-
bution, and Carbon/Nutrient Flow. Ph.D. Dissertation, University
of Georgia. 1 12 pp.
lamburri, M. N. 1990. Oyster larvae settle to waterborne chemical
factors released by adult conspecifics and by bacteria films. Am. Zool.
30: 97A.
Veitch, F. P., and H. Hidu. 1971. Gregarious setting in the American
oyster Crassosirea virginica Gmelin: I. Properties of a partially punned
"setting factor." Chesapeake Sci. 12: 173-178.
Walch, M., M. P. Labare, R. M. Weiner, R. R. Colwell, W. K. Fitt, and
D. B. Bonar. 1987. Use of specific bacterial bionlms and their prod-
ucts to enhance spat set of the oysters Crassosirea virginica and C.
gigas. J. Shell/ish Res 7: 1 79- 1 80.
Weiner, R. M., M. Walch. M. P. Labare, D. B. Bonar, and R. R. Colwell.
1989. Effect of biofilms of the marine bacterium Alleromonas col-
wettiana (LST) on set of the oysters Crassosirea gigas (Thunberg,
1793) and C virginiai {Gmelin, 1791). J Shellfish Res. 8: 117-123.
Wilkerson, F. P.. and R. K. Trench. 1986. Uptake of dissolved inorganic
nitrogen by the symbiotic clam Tridacna gigas and the coral Acroporu
sp. Mar Biol. 93: 237-246.
/.immer-Faust, R. K. 1990. Settlement behavior of larvae is revealed
using computer-video motion analysis. Am Zool 30: 98A.
Reference: Biol. Bull 182:409-415. (June, 1992)
Proline Synthesis During Osmotic Stress in Megalopa
Stage Larvae of the Blue Crab, Callinectes sapidus
RONALD S. BURTON
Program in Evolutionary Biology. Department of Biology. University of Houston,
Houston, Texas 77204-5513
Abstract. The free amino acid (FAA) pool of individual
Callinectes sapidus megalopas acclimated to 100% sea-
water averaged over 56% larger than that of 50% seawater
acclimated megalopas. Most of the difference was due to
a four-fold increase in proline concentration at the higher
salinity. In 100%- seawater, proline comprises 64% of the
total FAA pool in megalopas; this contrasts with the role
of proline in adult tissues where it never exceeds 25% of
the total FAA pool. Metabolic tracer studies using 14C-
glucose and l4C-glutamate as radiolabelled precursors
showed that dc novo synthesis of proline was very low
unless induced by hyperosmotic stress. The induction of
the synthetic pathway was inhibited by cycloheximide, a
protein synthesis inhibitor. These results suggest that the
induction of proline synthesis is regulated by the synthesis
of either one of the enzymes catalyzing the three steps in
the glutamate to proline pathway or a protein acting to
stimulate the activity of one of those enzymes.
Introduction
The adjustment of intracellular free amino acid (FAA)
concentrations plays an important role in acclimation to
salinity change in Crustacea and a diversity of other ma-
rine invertebrate taxa (see reviews by Florkin and Schof-
feniels, 1969;Gilles. 1975, 1979; Schoffeniels, 1976). High
intracellular FAA concentrations apparently function to
balance high inorganic ion concentrations in the hemo-
lymph of animals exposed to elevated salinity. Only a few
non-essential amino acids such as alanine, proline, and
glycine are major contributors to the response and show
rapid, quantitatively important, changes in concentration
following changes in environmental salinity. One or more
of these FAA typically obtain intracellular concentrations
Received 19 November 1991: accepted 10 March 1992.
in excess of 0. 1 M in seawater-acclimated Crustacea where
the total FAA pool may account for as much as 50% of
the total intracellular osmolyte pool (Bowlus and Somero,
1979).
While changes in FAA pool sizes have been widely
documented, relatively little is known about the regulation
of FAA concentrations during osmotic stress. The most
widely cited hypothesis involves the direct action of in-
organic ions on a key enzyme, glutamate dehydrogenase
(GDH, EC 1.4.1.2), which catalyzes the reductive ami-
nation of «-ketoglutarate to form glutamate (see Gilles,
1979; Gilles and Pequeux, 1983; Hochachka and Somero,
1984). Increasing medium salinity is postulated to result
in increases in intracellular NaCl concentrations that may
directly stimulate GDH activity, resulting in the synthesis
of glutamate. Because glutamate is the amino group donor
for synthesis of alanine and aspartate (and probably gly-
cine) and a direct precursor for proline, the increased glu-
tamate synthesis could drive, by mass action, the synthesis
of these other FAAs. Other key enzymes in FAA synthesis
are unaffected by changes in inorganic ions (e.g., trans-
aminases), while some involved in FAA catabolism are
inhibited by increased inorganic ion concentrations (e.g.,
serine hydrolyase). Combined, these effects are thought
to alter the synthesis/catabolism balance for FAA and re-
sult in their accumulation.
Unfortunately, while data continue to support the oc-
currence of de novo synthesis of FAA in response to hy-
perosmotic stress (e.g.. Burton, 1986), few data directly
support the above model for regulation of FAA synthesis.
The effect of NaCl and other inorganic ions on glutamate
dehydrogenase activity in vitro has proven to be complex
at best, and some investigators now propose that the major
change leading to FAA accumulation is not increased
synthesis, but rather reduced FAA catabolism (Gilles,
1979; Gilles and Pequeux, 1983). However, Burton ( 1986,
409
410
R. S. BURTON
199 la, b) has shown that in the euryhaline intertidal co-
pepod Tigriopus californicus, detectable proline synthesis
is observed only during hyperosmotic stress. After proline
accumulates for approximately 24 h, proline synthesis is
effectively turned off. This is in contrast to other FAAs
(such as alanine, glutamate, and aspartate), which are
synthesized continuously under a variety of salinity re-
gimes. Clearly, proline synthesis is not simply driven by
mass action following increased glutamate production
because glutamate production occurs under all salinity
conditions and glutamate pool size does not change
markedly during hyperosmotic stress. Similarly, regulation
of the proline pool cannot be the result of changes in
proline catabolism alone, because such a model could not
account for the fact that 14C-labelling of glutamate (from
labelled bicarbonate) occurs under constant salinity (50%
or 100% SW), while no labelling of proline is observed
under these conditions (Burton, 1986). These data are in
direct conflict with the mass action synthesis model dis-
cussed above.
Recently, by using in vivo translation inhibitor studies,
we have shown that the induction of proline synthesis in
T. californicus in response to hyperosmotic stress requires
protein synthesis (Burton, 199 Ib). By providing I4C-(U)-
L-glutamate as a proline precursor, evidence was obtained
that the ultimate site of action for protein synthesis in-
hibitors was in the three-step pathway between glutamate
and proline. This work suggests that hyperosmotic stress
induces the synthesis of one or more of the enzymes in
the glutamate to proline biosynthetic pathway or a protein
that stimulates the activity of these enzymes. Given that
this mechanism for the regulation of FAA metabolism
has not been previously documented among marine
Crustacea, it was of interest to determine the generality
of our T. californicus work by performing similar studies
on other, taxonomically distant, crustacean species.
Several criteria were important in choosing an appro-
priate study system for testing the mechanism of induction
of proline synthesis. First, the test organism should be a
euryhaline osmoconformer where adjustment of FAA
concentrations function in salinity acclimation. Second,
because proline is only a minor constituent of the FAA
pool in some species, a species was needed in which proline
was known to be an important contributor to the FAA
pool. Finally, for analytic convenience, we sought a small
organism because smaller quantities of tracer isotopes are
necessary for metabolic studies. One system meeting these
criteria is the blue crab, Callinectes sapidus. an abundant
portunid that experiences substantial salinity variation in
its natural estuanne habitat along the Texas coast. The
participation of FAA in osmotic acclimation of adult C.
sapidus has previously been studied (Gerard and Gilles,
1972; Engel, 1977), and proline was found to be a major
contributor to the osmolyte pool in each tissue studied.
While adult C. sapidus is too large for the in vivo radi-
otracer studies needed to address mechanisms of proline
synthesis, C. sapidus megalopas (dry weight of approxi-
mately 0.4 mg) are locally abundant and easily maintained
in the lab. In the work described below, osmotically in-
duced changes in FAA concentrations are documented
in C. sapidus megalopas and the role of protein synthesis
in regulating these changes is assessed via in vivo appli-
cation of the protein synthesis inhibitor cycloheximide.
Materials and Methods
C. sapidus megalopas were collected with a hand-pulled
beam trawl in shallow water ( < 1 . 5 meter) along the sandy
Gulf coast beach of Galveston Island, Texas, in early June
to August 1991. Ambient salinities ranged from 17 to 37
ppt. Animals were maintained at room temperature
(23°C) and acclimated for 3-5 days at 17 ppt (50% sea-
water = 50% SW) and 34 ppt (100% SW) before being
exposed to experimental treatments. Animals were fed
commercial flake fish food (Tetramin) during acclimation.
C. sapidus megalopas were initially identified by com-
paring them to the description presented in Costlow and
Bookout (1959). Numerous megalopas molted to the first
crab stage in our aquaria within a few days of capture;
these were identified as C. sapidus as described in Williams
(1984).
Procedures for studying the incorporation of labelled
precursors into the FAA pool in individuals of C. sapidus
were as follows: prior to exposure to precursor-laced me-
dium, animals were pretreated for 1 h with an antibiotic
mixture ("AM 4" of Provasoli el ai, 1959) in filtered (0.2
A/), buffered (30 mM HEPES) commercial (Instant Ocean)
artificial seawater (SW) of appropriate salinity. The effec-
tiveness of this antibiotic mixture in preventing contam-
inating bacterial growth was previously tested (Burton,
1991a). Radioactive precursor, l4C-(U)-L-glutamate,
(Sigma Chemical Company, 229.4 mCi/mmol) or I4C-
(U)-D-glucose, (Sigma Chemical Company, 255 mCi/
mmol) was added to a small volume (5 yuCi/150 ^1) of
medium of appropriate salinity; all media contained the
antibiotic mixture. Experimental treatments involving the
translation inhibitor cycloheximide also used a 1-h pre-
treatment period (with antibiotics and cycloheximide)
prior to salinity transfer. Transfers between pretreatment
and treatment media were carried out by pipetting indi-
vidual megalopas onto filter paper and then moving them
(with a fine forceps) into 1.5 ml microcentrifuge tubes
containing the desired treatment medium. Up to six me-
galopas were treated together in a single tube. Handling
was identical for controls and treatments and did not di-
rectly result in any mortality. Osmotic concentrations of
artificial SW solutions were routinely determined with a
hand refractometer and checked with a vapor pressure
osmometer (Wescor Model 5500).
REGULATION OF PROLINE SYNTHESIS
411
Following experimental exposures (typically 3-6 h),
animals were individually sacrificed and FAA extracted
in 100 n\ of 80% ethanol and then dried under vacuum.
Samples were resuspended in 60 n\ of 0.1 M sodium bi-
carbonate and then reacted with 40 n\ of dansyl chloride
in acetone (0.5 mg/ml) for 90 min at room temperature
to fluorescently label primary and secondary amino
groups. FAA analysis was carried out on dansyl derivatives
of the FAA using reverse-phase high pressure liquid chro-
matography (HPLC) (CIS "Hypersil" 5 /u 4.6 X 250 mm
cartridge column. Alltech Assoc.) with fluorescence de-
tection; peaks representing FAA were quantified with a
computing integrator and were individually collected di-
rectly into minivials for liquid scintillation counting (see
Burton, 1986, for further HPLC details). Although glycine,
taurine, alanine, and proline derivatives were completely
resolved, there was some difficulty in resolving dansyl-
glutamate from dansyl-aspartate; data for the combined
glutamate/aspartate peak are presented here as "gluta-
mate." To determine whether the confounding of gluta-
mate and aspartate would have a significant effect on es-
timates of glutamate specific activity, one sample from
each experimental treatment was analyzed by one-di-
mensional thin-layer chromatography (TLC), as follows.
Three samples and one lane of dansyl-FAA standards
(Sigma Chemical) were spotted in four lanes on a 5 X 20
cm polyamide 6 TLC plate (Baker Chemical). Chro-
matograms were run in a chloroform-t-amyl alcohol-acetic
acid (70:30:3) solvent system until the solvent front mi-
grated 15 cm. Glutamate and aspartate pool sizes were
qualitatively assessed under UV illumination, and distri-
bution of radiolabel was determined by a 48-72-h auto-
radiographic exposure. In most cases, the aspartate spot
was too faint to be detected by eye. Subsequently, chro-
matographic regions in the sample lanes corresponding
to aspartate and glutamate standards were cut out and
eluted for scintillation counting. In all tests, the bulk of
the label (minimum 75%) was recovered in the glutamate
region. Because none of our qualitative results are signif-
icantly affected by reducing the counts recovered in glu-
tamate by such a factor, we concluded that pooling the
glutamate and aspartate peaks via HPLC did not introduce
significant error into the results presented here.
Levels of FAA measured by HPLC are presented here
in units of nanomoles/larva. The mean (±S.E.) wet weight
of a larva was 1 .37 ± 0.04 mg; dry weight was 0.39 ± 0.02
mg. Although the reported values can, therefore, be con-
verted to more common units (e.g., mmoles/kg tissue wa-
ter, or mmoles/g dry weight), the fact that whole mega-
lopas were homogenized would make it difficult to com-
pare the values presented here to values reported for adult
tissues. This is because our wet weights include gut water
content, and our dry weights consist primarily of exo-
skeleton rather than actual FAA-containing tissue.
Results
FAA pool o/Ca!linectes sapidus megalopas acclimated
to 50% and 100% SW
Following collection from ambient 100% SW, groups
of megalopas were acclimated to 100%. and 50% SW for
five days as described above. Results of FAA analyses are
shown in Figure 1. In 50% SW, taurine and glycine are
the dominant FAAs, comprising approximately 42% and
25% of the measured pool, respectively. The total FAA
pool of 100% SW acclimated animals averaged over 56%
larger than that of 50%< SW acclimated animals (one-tailed
/-test, P = 0.022). The only amino acid that contributed
significantly to the increased pool was proline (taurine
and glutamate actually showed relatively minor but sta-
tistically significant decreases); in 100% SW, proline com-
prised over 64% of the FAA pool.
Incorporation of radiolabelled glucose into the FAA pool
Megalopas were presented with 14C-(U)-D-glucose (5
jiCi/150 i/l of medium) under three salinity treatments:
constant 50% SW (involved transfer between media of
the same salinity), constant 100%. SW, and immediately
following hyperosmotic transfer from 50% to 100% SW.
Larvae were sampled at two time points: 3 h and 6 h after
treatment. After 3 h of hyperosmotic stress, concentrations
of the five FAAs measured had not increased significantly
above the 50% SW control (taurine showed a small but
statistically significant drop). Although FAA concentra-
tions had not yet changed, analysis of radiotracer incor-
poration shows evidence of significant changes in FAA
metabolism. Although 80-90% of all recovered radioac-
tivity in FAA is in the alanine pool under each salinity
treatment, only proline showed significant variation in
specific activity among treatments (Fig. 2). Under either
constant salinity treatment, proline specific activity av-
1
I
<u
I
I
GLU
f .
GLY
1'
7
'AU
I
ALA
* •
PRO
j
50 100 50 WO 50 100 SO WO 50 100
Acclimation Salinity (% Seawater)
Figure 1 . FAA concentrations in individual Callinectes sapidus me-
galopas acclimated to 50% and 100% SW for five days following collection
from ambient 100% SW. Error bars are 95% confidence intervals (n
= 5 and 6 to the two treatments, respectively).
412
R. S. BURTON
40'
GLUTAMATE
"55
T
| «.
•o; 'S
5| '".
i
{
<
? "'
c
" 20
I
A «o
50% 50%-100% 100%
ALANINE
PROLINE
50% 50%- 100% 100%
SALINITY TREATMENT
50% 50%-100% 100%
° fhrr
Figure 2. FAA concentrations and specific activities in individual C W/mcr/c.s Mpiilu<, megalopas subjected
to three treatments: acclimated to 50% and transferred to 50% SW or 100% SW. or acclimated to [00% SW
and transferred to 100% SW. l4C-(U)-D-glucose was added and animals were sampled at 3 and 6 h. Four
to six individuals were analyzed per treatment. Error bars are ±1 S.E.M. Note scale differences among panels.
eraged less than 2 dpm/nmole while that of glutamate. a
direct proline precursor, averaged 95 dpm/nmole, nearly
two orders of magnitude higher (note that because aspar-
tate was lumped with glutamate and all counts appear to
be in glutamate, this specific activity is an underestimate
of true glutamate activity); this fact indicates that proline
is essentially not being synthesized dc novo from glucose
carbon under constant salinity conditions. In contrast,
proline specific activity in the hyperosmotic stress treat-
ment increased dramatically (to approximately 50% of
the glutamate specific activity in the same treatment), in-
dicating the induction of proline biosynthesis by the hy-
perosmotic stress treatment. Hence, although glutamate
and alanine specific activities were not significantly influ-
enced by the salinity treatments, proline specific activity
increased by two orders of magnitude within 3 h of hy-
perosmotic stress. Glycine and taurine showed no label
incorporation in any treatment and are omitted from Fig-
ure 2.
By 6 h. a significant concentration increase was detected
for alanine in the stress treatment over 50% SW controls;
while mean proline concentration increased by a factor
of four, inter-individual variance was large and the mean
difference was not statistically significant (Fig. 2). Again
only the specific activity of proline was elevated over con-
stant salinity controls.
Incorporation of radiolabelled glutamate into the f-'AA
pool
Megalopas were presented with l4C-(U)-L-glutamate
(5 ^Ci/150 ^1 of medium) under both constant salinity
(100% SW) and hyperosmotic stress conditions (50-
100%. SW transfer) to further ascertain that induction
of proline synthesis involves the regulation of the glu-
tamate to proline pathway. The high salinity control
was employed because if proline is being synthesized
under constant salinity conditions, it should be most
evident in high salinity where proline pool sizes are
large. Results are presented in Figure 3. Although the
size of the glutamate pool was slightly larger in the con-
trol versus the osmotic stress conditions (two-tailed t-
test. P < 0.01 ). radiolabel recovered in glutamate and
glutamate specific activity did not differ. This indicates
that a comparable pool of labelled glutamate was avail-
able for proline synthesis under both sets of conditions
(if anything, slightly more glutamate was available un-
der the control conditions). In contrast, even though
the proline pool is significantly larger in control animals
(P < 0.02), label recovered in the pool and proline spe-
cific activity is lower under control conditions (P <
0.001 for each measure). Hence, while some label is
observed in proline under constant salinity conditions,
the near-zero specific activity suggests that the flux from
glutamate to proline under these conditions is very low.
In fact, if glutamate is the primary precursor of the pro-
line pool, a flux from glutamate to proline should lead
to equilibration of the specific activities of the two pools.
Paired /-tests show that glutamate and proline have dif-
ferent specific activities (within each individual) under
control conditions (P < 0.01 ), but not following the 4-
h hyperosmotic stress treatment (P > 0. 1 ).
REGULATION OF PROLINE SYNTHESIS
413
GLUTAMATE
ALANINE
PROLINE
1
:| 60'
\
'<; 40-
o>
'o
(
E
fi 20 '
^
9 a
• *
LL 0"
^5 6000
1 *
3
|
•E 4000
, ,
s
Q)
0 POOO
1
1
e
. if
"5 800
9' ' X
^ 600
.g.
^ 400
^ 200
\
t
1 °
%
s % s « 1 %>
^ § I g 1 S
0 CO § CO CJ ^
TREATMENT
Figure 3. FAA concentrations and specific activities in individual
Callinecies sapidiis megalopas acclimated to 100"^ SW and transferred
to 100% SW (Control) and acclimated to 50% SW and transferred to
100% SW (Stress) for 4 h in the presence of l4C-(U)-L-glutamate. Error
bars are 95% confidence intervals (n = 6 individuals per treatment).
Effects of the protein synthesis inhibitor cycloheximide
on FAA synthesis
Because the results above clearly demonstrate the in-
duction of proline synthesis from glutamate during re-
sponse to hyperosmotic stress, we used cycloheximide
(CHX) as a protein synthesis inhibitor to address the role
of protein synthesis in this induction. Paired groups of
megalopas were exposed to hyperosmotic stress, with one
group being treated with CHX. Results are shown in Fig-
ure 4. Levels of glutamate, labelling of the glutamate pool.
and glutamate specific activity do not differ between con-
trol and CHX treatments. In contrast, all three measures
of proline were affected. Although the size of the proline
pool was only slightly decreased by CHX (one-tailed t-
test, P < 0.05), both label recovered in proline and proline
specific activity were dramatically reduced by CHX treat-
ment (P < 0.005). Interestingly, alanine pool sizes in-
creased by over 40% in the CHX treatment (P < 0.05).
Neither glycine nor taurine pool sizes were significantly
influenced by CHX treatment (P > 0.25, data not shown).
The magnitude of this change is not enough to compensate
for reduced proline synthesis, but it does indicate that
£
40
30
^
3> 20
O
t w
GLUTAMATE
ALANINE
Pt
1OLINE
^
^ \
$ •
* *
I
"6
I
6000
4000
2000
I
J L
"a? 800
"5
E
c 60°
1
^ 400
S
'
<
^ 200
.y
$
'Q
* »
- •
CO"
Control CHX
Control CHX
TREATMENT
Control CHX
Figure 4. FAA concentrations and specific activities in individual
Caltinectes sapuliis megalopas acclimated to 50% SW for 3 days and
transferred to 100% SW for 4 h in the presence of l4C-(U)-L-glutamate.
CHX groups were pretreated with 10 m.V cycloheximide for 1 h. Error
bars are 95% confidence intervals (n = 6 individuals per treatment).
414
R. S. BURTON
CHX treatment was specifically inhibitory to proline (but
not alanine) synthesis.
Discussion
Although the role of proline accumulation in the hy-
perosmotic response is variable among the Crustacea, such
an accumulation of proline is wide-spread among taxa,
having been observed among bacteria (Le Rudulier et al,
1984), fungi (Ho and Miller, 1978), and metaphytes (Bog-
gess et ai, 1976), as well as among marine invertebrates
(FlorkinandSchoffeniels, 1969;Gilles, 1975, 1979;Schof-
feniels, 1976). The mechanisms underlying proline ac-
cumulation may include protein degradation, uptake from
the medium, and de novo synthesis. The role of each
mechanism varies among taxa: for example, among mi-
croorganisms, gram-positive bacteria appear to regulate
the synthesis or degradation of proline. whereas gram-
negative bacteria achieve accumulation primarily via up-
take from the medium (Csonka, 1989).
There is substantial variation among species with regard
to the importance of proline in the FAA pool during hy-
perosmotic response (Claybrook, 1983). Among the
Crustacea, proline is insignificant in the FAA pools of
some species but the dominant contributor to the FAA
pool in others. Furthermore, different tissue types vary
dramatically in composition of the FAA pool. While pro-
line is a major contributor to the pool in most adult C.
sapidus tissues (Gerard and Gilles, 1972), it never ac-
counted for more than 25% of the FAA pool in seawater-
acclimated animals. In contrast, our data indicate that
proline is the predominant FAA in megalopas, comprising
over 50% of the FAA pool. Whether such ontogenetic
changes in the composition of the FAA pool are common
to other Crustacea has yet to be studied.
The regulation of FAA metabolism in response to os-
motic stress among the Crustacea and other marine in-
vertebrates is poorly understood. As discussed above and
in Burton (199 la, b), models of direct inorganic ion effects
on specific enzymes in FAA metabolism appear to be in-
sufficient to explain the regulation of proline synthesis for
two reasons. (1) Patterns of incorporation of radioactive
precursors into proline indicate that rate of proline syn-
thesis is nearly undetectable unless induced by hyperos-
motic stress (Burton, 1986). (2) Because l4C-labelled glu-
tamate was provided as a precursor and inhibition of pro-
tein synthesis prevented proline synthesis which occurred
in the absence of inhibitor, we can conclude that protein
synthesis inhibition acts somewhere in the glutamate to
proline pathway. Based on available information from
bacteria and yeast, three gene loci encode the enzymes:
7-glutamyl kinase, -y-giutamyl phosphate reductase, and
pyrroline-5-carboxylate reductase (Hayzer and Leisinger,
1980; Tomenchok and Brandriss, 1987). Although work
on the proline biosynthetic pathway has not yet progressed
to the genetic level among metazoans, available data sug-
gests that homologous gene-enzyme systems are present
(Smith et ai. 1980; Wakabayashi and Jones, 1983).
The results presented here for C. sapidus megalopas
are similar to those obtained by Burton (1986, 1991b) for
the copepod Tigriopus californicus in suggesting that pro-
line synthesis is specifically induced by increases in en-
vironmental salinity rather than simply driven by changes
in the synthetic rate of a precursor (i.e., glutamate). As in
the T. californicus system, induction of proline synthesis
appears to be dependant on protein synthesis. In both
systems, the inhibition of protein synthesis with cyclo-
heximide decreased proline synthesis and accumulation
during hyperosmotic stress but significantly increased al-
anine accumulation. Two possible explanations for the
enhanced alanine accumulation follow: ( 1 ) By directly
preventing the incorporation of alanine and other amino
acids into protein, cycloheximide might lead to measur-
able increases in components of the FAA pool. (2) By
preventing the induction of proline synthesis, cyclohexi-
mide increases the availability of alanine precursors (e.g.,
glutamate), thereby stimulating alanine synthesis. In T.
californicus, cycloheximide treatment resulted in signifi-
cant incorporation of l4C-label from glutamate into ala-
nine (presumably via glutamate catabolism to malate and
then pyruvate, a direct alanine precursor), supporting the
latter explanation. This effect was not observed in C. sap-
idus. so the validity of the two hypotheses cannot be re-
solved with the data available. It should be noted, however,
that of the five FAA monitored (glutamate, glycine, tau-
rine, alanine, and proline), only alanine showed increased
levels in response to cycloheximide treatment. This sug-
gests that the former hypothesis alone is unlikely to ac-
count for the observed pattern of FAA accumulation when
protein synthesis is inhibited.
The similarities between proline regulation in C. sap-
idus megalopas and T. californicus suggest that the in-
duction of proline synthesis by hyperosmotic stress might
be a common regulatory mechanism among the Crusta-
cea. Protein synthesis is clearly required for the induction
of proline synthesis in both species. While one must be
cautious about generalizing on the basis of only two spe-
cies, our results suggest a need for molecular tools to de-
termine if the responsible protein is an enzyme in the
pathway itself or a regulatory protein of some sort that
stimulates existing enzymes to initiate proline synthesis.
Acknowledgments
I thank L. Kordos, J. Bishop, and H. Nguyen for col-
lecting the crab megalopas, and H. Nguyen for technical
assistance. J. Bishop and two anonymous reviewers pro-
vided thoughtful comments on the manuscript. This work
REGULATION OF PROLINE SYNTHESIS
415
was supported by Texas Sea Grant. NSF Grant DCB-
881 1227. and the University of Houston Coastal Center.
Literature Cited
Boggess, S. F., C. R. Stewart, D. Aspinall, and L. G. Paleg. 1976. Effect
of water stress on proline synthesis from radioactive precursors. Plant
Physiol. 58: 398-401.
Bowlus, R. D.. and G. N. Somero. 1979. Solute compatibility with
enzyme function and structure: rationales for the selection of osmotic
agents and end-products of anaerobic metabolism in marine inver-
tebrates. / Exp. Zool 208: 137-152.
Burton, R. S. 1986. Incorporation of l4C-bicarbonate into the free ammo
acid pool during hyperosmotic stress in an intertidal copepod. J.
Exp. Zool. 238: 55-61.
Burton, R. S. 1991a. Regulation of proline synthesis during osmotic
stress in the copepod Tigriopits californicus. J. Exp Zool. 259: 166-
173.
Burton. R. S. 1991b. Regulation of proline synthesis in osmotic re-
sponse: effects of protein synthesis inhibitors. J. Exp. Zool. 259: 272-
277.
Cla> brook, D. L. 1983. Nitrogen metabolism. Pp. 163-2 1 3 in The Bi-
ologv of Crustacea. Vol. 5. L. H. Mantel, ed. Academic Press, New
York.
Costlow, J. D., Jr., and C. G. Bookout. 1959. The larval development
of Callineaes sapidiis Rathbun reared in the laboratory. Biol. Bull.
116: 373-396.
Csonka, L. N. 1989. Physiological and genetic responses of bacteria to
osmotic stress. Microbiol. Rev. 53: 121-147.
Engel, D. W. 1977. Comparison of the osmoregulatory capabilities of
two portunid crabs, Callineaes sapidits and C. similis. Mar. Biol.
41: 275-279.
Florkin, M., and E. Schoffeniels. 1969. Molecular Approaches to Ecol-
ogy. Academic Press. New York. 203 pp.
Gerard, J. F., and R. Gilles. 1972. The free amino acid pool in Cal-
linectes sapidits (Rathbun) tissues and its role in the osmotic mtra-
cellular regulation. J. Exp. Afar. Biol. Ecol. 10: 125-136.
Gilles, R. 1975. Mechanisms of ion and osmoregulation. Pp. 259-347
in Marine Ecology, I'o/. 2. Part I. O. Kinne, ed. John Wiley and
Sons, London.
Gilles, R. 1979. Intracellular organic osmotic effectors. Pp. 1 1 1-154
in Mechanisms of Osmoregulation in Animals, R. Gilles. ed. John
Wiley and Sons, New York.
Gilles, R., and A. Pequeux. 1983. Interactions of chemical and osmotic
regulation with the environment. Pp. 109-177 in The Biology of
Crustacea. I 'ol. 8. F. J. Vernberg and W. B. Vernberg, eds. Academic
Press, New York.
Hayzer, D. J., and Th. Leisinger. 1980. The gene-enzyme systems of
proline biosynthesis in Escherichia coli. J. Gen. Microbiol. 118: 287-
293.
Ho, K. H., and J. J. Miller. 1978. Free proline content and sensitivity
to desiccation and heat during yeast sporulation and spore germi-
nation. Can. J Microbiol. 24: 312-320.
Hochachka, P. W., and G. N. Somero. 1984. Biochemical Adaptation.
Princeton University Press. Princeton. 537 pp.
Le Rudulier, D., A. R. Strom, A. M. Dandekar, L. T. Smith, and R. C.
Valentine. 1984. Molecular biology of osmoregulation. Science 224:
1064-1068.
Provasoli, L., K. Shiraishi, and J. R. Lance. 1959. Nutritional idio-
syncrasies of Artemia and Tigriopits in monoxenic culture. Annals
N. Y.Acad.Sci.n-.25Q-l(>l.
Schoffeniels, E. 1976. Adaptations with respect to salinity. Biochem.
Soc. Symp. 41: 179-204.
Smith, R. J., S. J. Downing, J. M. Phang, R. F. Lodato, and T. T. Aoki.
1980. Pyrroline-5-carboxylate synthase activity in mammalian cells.
Proc. Natl. Acad. Sci. USA 77: 5221-5225.
Tomenchok, D. M., and M. J. Brandriss. 1987. Gene-enzyme rela-
tionships in the proline biosynthetic pathway of Saccharomyces cer-
evisiae. J. Bactenol. 169: 5364-5372.
Wakabayashi, Y., and M. E. Jones. 1983. Pyrroline-5-carboxylate
synthesis from glutamate by rat intestinal mucosa. J. Biol. Chem.
258: 3865-3872.
Williams, A. B. 1984. Shrimps, Lobsters, and Crabs of the Atlantic
Coast of the Eastern United States. Maine to Florida. Smithsonian
Institution Press. Washington. DC. 550 pp.
Reference: Biol Bull 182: 416-423. (June, 1992)
Behavioral Regulation of Hemolymph Osmolarity
Through Selective Drinking in Land Crabs,
Birgus latro and Gecarcoidea lalandii
CHRISTIAN A. COMBS, NICOLE ALFORD, ANGELA BOYNTON,
MARK DVORNAK, AND RAYMOND P. HENRY
Department of Zoology and Wildlife Science. 101 Cary Hall, Anbwn University, Alabama 36849
Abstract. Drinking behavior in Birgus latro and Gecar-
coidea lalandii was videotaped under controlled labora-
tory conditions. B. latro displayed the drinking behavior
typically observed in nature, spooning up water with the
chelae (Lister, 1888; Gross, 1955). G. lalandii is docu-
mented for the first time displaying this same behavior;
however it obtained water primarily through immersion.
Under normal hydrated conditions (hemolymph osmo-
larities < 1050 mOsm) B. latro showed no preference for
drinking fresh or seawater. When dehydrated (hemolymph
osmolarities > 1050 mOsm) B. latro altered its drinking
behavior and showed a distinct preference for freshwater.
This strategy resulted in restoration of original hemo-
lymph osmolarities and wet weights and was accomplished
through periods of intensive drinking activity. Conversely.
G. lalandii never experienced true dehydration; rather,
the hemolymph became hyperosmotic compared with
control animals. This species preferred freshwater both
under normal and hemoconcentrated conditions. G. la-
landii was also able to osmoregulate behaviorally and was
able to restore hemolymph osmolarities to normal con-
centrations via immersion in freshwater following exper-
imentally induced hemoconcentration. Possible physio-
logical and ecological reasons for the differences in water
uptake strategies and preferences are discussed.
Introduction
The transition from the aquatic to the terrestrial en-
vironment was both problematic and beneficial for crabs.
Although oxygen was more readily available and there
were new resources to exploit, certain morphological.
Received 9 September 1991; accepted 26 February 1992.
physiological, and behavioral strategies were required to
overcome the barrier of desiccation and concomitant in-
crease in hemolymph ion concentrations. Changes in he-
molymph ion concentrations can profoundly affect many
physiological processes in crustaceans including respira-
tion, acid-base status, intracellular fluid volume, nitrog-
enous waste elimination, and enzyme function (Harris,
1977; Burrgren and McMahon 1981, 1988; Morris el al.,
1988; Wheatly et al.. 1984; Wood el al. 1986; for reviews
see Huggins and Munday. 1968: Schoffeniels, 1976; Gilles
andPequeux. 1981; Mangum, 1981; Taylor, 1982; Yan-
cey et al.. 1982). In land crabs, dehydration and changes
in hemolymph concentration are resisted using combi-
nations of adaptations such as immersion, burrowing,
water storage in the body or branchial chambers, evolu-
tionary reduction in gill size, urine reprocessing, excretion
of nitrogenous waste as urea or uric acid, and drinking
(Bliss and Mantel, 1968; Bliss, 1979; Mantel and Farmer,
1983; Powers and Bliss, 1983; Wolcott and Wolcott, 1985,
1991; Greenaway, 1988; Greenaway et al.. 1988; T. G.
Wolcott. 1988; D. L. Wolcott. 1991).
Drinking, or spooning up water with the chelae to the
mouthparts, has been documented in Gecarcoidea natalis.
Geograpsus grayi. Cardisoma guanhiimi (Gross et al.,
1966), C. carnifex (Greenaway, 1988), Gecarcoidea lal-
andii (this paper), and in Birgus latro (Lister, 1888). Fresh
and seawater are available to most land crabs, but it is
not known precisely how the more terrestrial species use
the two to regulate the concentration of their hemolymph.
These crabs inhabit islands in the Indo-Pacific region
where rainfall is seasonal and water sources other than
the ocean may become scarce at certain times of year
(Gross, 1964). This study examines the drinking prefer-
ence of two of the more strictly terrestrial crabs, Birgus
416
BEHAVIORAL OSMOREGULATION IN CRABS
417
latro (Anomura) and Gecarcoidea lalandii (Brachyura),
when hemolymph concentrations are normal and hemo-
concentrated (with possible dehydration).
Materials and Methods
Animal collection and maintenance
Specimens of both species, Birgus latro (500-2100 g)
and Gecarcoidea lalandii (65-220 g) were obtained from
the islands of Palau and Pohnpei and were shipped via
air-freight in damp burlap inside coolers or ice chests.
They were maintained in isolated Nalgene tanks (B. latro)
or fabricated wooden pens with plexiglass partitions (G.
lalandii) in a dark room at approximately 25 °C. All were
fed coconut, lettuce, and apples even other day, and were
given either dechlorinated tapwater or seawater (35-40
ppt, as determined with a refractometer) depending on
the subsequent testing regime.
Protocol
Observation chambers were constructed using standard
75 and 1151 aquaria fitted with a false bottom of plexiglass
to facilitate viewing and minimize water spillage. Two
circular holes were cut side-by-side in the false bottom to
allow access to a pair of glass preparation dishes (115
X 50 mm. Fisher) that were secured with silicone to the
aquarium floor. The tops of the dishes were flush with
the plexiglass platform, simulating pools of water. The
chamber was darkened on three sides to isolate each crab.
The uncovered end of the aquaria permitted head-on
viewing of both bowls. A plexiglass cover with air holes
for ventilation was secured to each tank to prevent escape.
Drinking behavior of animals was recorded for indi-
viduals displaying hemolymph osmotic concentrations
typically found for crabs sampled in the field (<1050
mOsm) (as reported by Henry and Cameron, 1981;
Greenaway, 1988) and after hemoconcentration with
possible dehydration (hemolymph osmolarity > 1050
mOsm). The former condition was maintained by allow-
ing crabs to drink ad libitum from both fresh and seawater.
Hemoconcentration was achieved through a combination
of water deprivation and allowing access to only hyper-
saline water (>35 ppt). Animals were weighed daily and
were not allowed to lose more than 12% of their initial
wet weight, a value that was within the maximal tolerable
levels of dehydration as reported by Kormanik and Harris
( 1 98 1 ) and Burggren and McMahon (1981). Hemolymph
osmolality was measured concurrently to determine at
what point values exceeded 1050 mOsm, after which an
experiment was begun.
Individual crabs were randomly assigned their initial
condition, normal or hemoconcentrated. Behavior of each
individual was recorded under both normal and hemo-
concentrated conditions, allowing at least a 48 h recovery
period between experiments.
All crabs were allowed to adjust to the chambers for
24 h prior to observation/videotaping, while remaining
on their water regime. Immediately preceding an obser-
vation period, crabs were weighed to the nearest 0. 1 g on
a top loading balance (Sartorius), and a blood sample (0. 1
ml) was taken from the infrabranchial sinus. The water
in each chamber was replaced with 225 ml of seawater
(35 ppt) and 225 ml of fresh (deionized) water was placed
randomly in either the right or left dish. The crabs were
then recorded for 12 h (1800-0600 h) using a Panasonic
VHS Recorder Model AG-HT3. After 1 2 h, the crabs were
weighed, blood samples were drawn, and the water volume
remaining in each bowl was measured to the nearest 1.0
ml. Evaporative water loss was quantified using duplicate
bowls of freshwater and seawater placed in an empty
chamber. Temperature and relative humidity were also
recorded using a ExTech Instruments Digital Humidity/
Temperature Meter.
Hemolymph samples were placed in microcentrifuge
tubes and kept on ice. After being sonicated at 20 watts
for 10 s with a Microson Cell Disrupter CM-1 converter
(Heat Systems — Ultrasonics, Inc.), the samples were cen-
trifuged for 1 min using a Micro-Centrifuge Model 235B
(Fisher). Osmolarity was determined on 10 n\ samples of
serum using a Wescor 5 100C Vapor Pressure Osmometer.
Quantification of the number of drinks, time spent
drinking, and time spent immersed in the water bowl were
determined from video tape analysis. An individual drink
was considered to be a cheliped sweep from the water to
the mouth and subsequent sweep by the maxillae over
the cheliped to remove the water (Lister, 1888). Time
spent drinking was designated as the time of the first che-
liped sweep to the time of the last cheliped sweep. Im-
mersion time was considered to be the amount of time
that a crab had part of its carapace submerged in a water
bowl.
Statistics
Paired /-tests were used to determine if there were dif-
ferences between water preferences within hemolymph
concentration treatments, weight differences between he-
molymph concentrations, and starting and ending he-
molymph concentrations between treatments. Analysis
of variance (ANOVA) was used to determine if there were
differences between the preferences of crabs for drinking
freshwater or seawater between hemolymph concentration
conditions in each species. To reduce biases inherent in
individual animals, only individuals that were tested at
both hemolymph conditions were included in the statis-
tical analysis. Scheffe's multiple comparisons test was used
to compare means of the variables between normal and
418
C. A. COMBS ET AL.
hemoconcentrated treatments. All data were tested for
normality using the Wilkes-Shapiro test and can be as-
sumed to be normally distributed unless specified. All sta-
tistical analyses were accomplished with the SAS™ sta-
tistical computer package (SAS Inst., Inc., 1982).
Results
Strategies of water uptake
The two species employed different strategies of water
uptake both to maintain a normal hemolymph osmotic
condition and to reduce hemolymph concentrations in
response to hemoconcentration (often accompanied by
dehydration in B. latro). B. latro used cheliped drinking
as the only means of water uptake, spending 100% of
their drinking time in that behavior; immersion in the
water bowls was not employed, although the bowls could
have accommodated at least portions of their bodies. The
observed drinking behavior was virtually identical to that
reported previously by Lister (1888). G. lalandii, however,
was observed in cheliped drinking behavior only 2% of
the time; the remainder of the time in contact with water
was spent with all or part of the carapace immersed in
the water bowl. When this species did engage in cheliped
drinking, the behavioral pattern was similar to that seen
in B. latro.
Normal water uptake and response to
hemoconcentration
All specimens of B latro that began an experiment in
a normal hemolymph concentration state (848 ± 22
mOsm) were able to maintain that state over the 12 h
observation period (Mest: P > 0.8 1 ) (Fig. 1 ). Under normal
conditions, B. latro spent an average of 6 1 min engaged
in drinking, performing 660 individual cycles of cheliped
sweeps during a 1 2-h experiment. This species showed no
preference for fresh or seawater, either with respect to the
percent of the total drinks taken from each bowl (/-test:
P> 0.73), the time spent drinking from each bowl (/-test:
P > 0.2), or the percent of volume that was drunk (/-test:
P> 0.2) (Fig. 2).
When individuals were hemoconcentrated prior to an
experiment (1171 ±51 mOsm), both the overall drinking
behavior and the drinking preference were altered. Ani-
mals in a hemoconcentrated condition increased their to-
tal drinking time by over four-fold to 273 min, taking an
average of 2702 individual drinks during that time. These
animals displayed a distinct preference for freshwater for
all three variables measured: total drinks, time of drinking,
and volume consumed (Mest: P < 0.01) (Fig. 2). In ad-
dition, the drinking preferences were different between
the two osmotic states (ANOV A: P < 0.0 1 ) with all control
crabs exhibiting the same preference and all hemocon-
centrated crabs exhibiting the same preference (Scheffe:
P < 0.05). This behavior led to a significant difference in
osmolarity changes before and after the 1 2-h tests between
the normal and hemoconcentrated treatments (/-test: P
< 0.02), with the hemoconcentrated animals reducing
their average osmolarity by 18% to 948 ± 26 mOsm. In
addition, differences in weight before and after the 12-h
tests were significantly different between control and
hemoconcentrated animals (/-test: P < 0.0 1 ) with normal
animals averaging only a 5 ± 12 g weight gain while he-
moconcentrated animals gained 75 ± 16 g (Fig. 1). There-
fore, B. latro compensated for hemoconcentration via
water gain, indicating that initial hemoconcentration was
accompanied by dehydration.
All specimens of Gecarcoidea lalandii that began an
experiment in a normal hemolymph concentration state
(947 ± 31 mOsm) were also able to maintain that state
over the 12-h observation period (/-test: P > 0.77) (Fig.
3). On average, specimens of G. lalandii that began an
experiment in a hemoconcentrated state (1168 ± 27
mOsm) were able to reduce their hemolymph concentra-
tion back to control levels (/-test: P > 0.30) over the 12-
h observation period. Freshwater was significantly pre-
ferred over seawater at both osmotic states (/-test: P
< 0.05 and P < 0. 1, respectively) when immersion time
was used as an indicator of behavioral osmoregulation
(Fig. 4). Moreover, there is a significant difference in the
amount of time spent immersed in seawater between the
two osmotic states, with more time being spent in seawater
when individuals started the experiments hemoconcen-
trated (ANOVA: P < 0.05, Scheffe: P < 0.05). The data
for number of cheliped drinks and time spent cheliped
drinking were not normally distributed, but freshwater
was overwhelmingly preferred at both hydration states
(Fig. 4) for the small amount of time, 2%, this species
spent cheliped drinking. End-volume differences in the
drinking bowls were not examined in this species due to
the propensity of the animals to splash water out of the
bowls while entering and exiting from them. This obscured
any differences that might have been due to their drinking
or absorbing water. Differences in weight before and after
the 12-h tests were not significantly different between
control and hemoconcentrated animals (/-test: P > 0.25).
Therefore, G. lalandii did not compensate for hemocon-
centration by water gain but rather appeared to reduce
hemolymph ion concentrations through ion exchange via
immersion in the ambient water.
Discussion
Both Birgus latro and Gecarcoidea lalandii can os-
moregulate behaviorally, by selecting drinking water of
the appropriate salinity, under laboratory conditions. Us-
ing this strategy they were able to maintain hemolymph
1450
-~ 1J50
£ 1150
€5*
X
i 1050
0 850
UJ
1 .50
750
650
BEHAVIORAL OSMOREGULATION IN CRABS
A DEHYDRATED
419
HYDRATED
PH PH BO NA B5 Bl MA R2 B3 83
PH BO B5 Bl NC Bl B3
HYDRATED
DEHYDRATED
2200 -
B
-,
T\
2000 •
</)" lg°° -
%
'/
S
—
|
^
< 1(00 -
^
J
/
K
J
y
2- 1400 -
^
p
i
$
;/•
/,
.
tk 1200 -
?
|
Jj
•>
'
^ 1000 -
P
/
|
/
•
^
/>
-
y
too
\
',
/
'',
-
7
%
^
^
/
r/
600 -
^
J
|
/
!
7|
T~
400 -
i
1
^
/
;
1
/
^
ri
.
200
!
:
\
^
1
,
1 1 .1
\
PH PH BO HA BS Bl NA B2 B) B)
PH BO B5 Bl NC Bl B3
Figure 1 . (A) Hemolymph concentrations (mOsm), and (B) weight (g), before (open bars) and after (shaded
bars) 1 2-hour tests involving Birguslatro. Left-hand regions of figure are tests with initially normal hemolymph
concentrations (<1050 mOsm) and nght-hand regions are tests with initially concentrated hemolymph
osmolanties (> 1050 mOsm). X-axis labels identify individual specimens.
concentration and wet weight at normal hydrated levels
during 12-h test periods. When hemoconcentrated, both
species could reduce their hemolymph osmolanties to
normal levels. This was accomplished by ion exchange in
G. lalandii, whereas B. latro took on water to dilute the
hemolymph.
Gross (1955) first documented the ability of B. latro to
osmoregulate behaviorally. The present study indicated
somewhat different results for drinking preference than
did Gross (1955). In our study, specimens of B. latro at
normal hemolymph concentrations (hypoosmotic to sea-
water) showed no preference for either fresh or seawaten
rather, they precisely regulated their hemolymph osmo-
larity through piecemeal drinking of both water types. In
contrast. Gross (1955) reported that hydrated crabs pre-
ferred freshwater, but did drink some seawater. Our results
coincide with those of Gross (1955) in that dehydrated
crabs (hemolymph concentrations hyperosmotic to sea-
water) preferred freshwater. The difference in results be-
tween the two studies may be a result of the manner by
420
C. A. COMBS ET AL
I
i
i
DEHYDRATED
HYDRATED
i
f
fw sw
HYDRATED
*
fw sw
DEHYDRATED
•3E-
fw
sw
- c
fw sw
DEHYDRATED
HYDRATED
f
i
i
i
fw sw fw sw
Figure 2. Drinking preference for freshwater or seawater according to
initial hemolymph concentration (hydrated and dehydrated) during 12-
hour tests involving Birgus lalro(n = 5). Drinking preference is expressed
as (A) mean percentage of total drinks. (B) mean percentage of total
time spent drinking, (C) mean percentage of total volume change from
water bowls (normalized for evaporative water loss).
which the quantification of the drinking preference was
obtained. Gross ( 1955) quantified drinking behavior in-
directly. Drinking behavior was monitored by etchings
made on Kymograph drums caused by depressions of
platforms over drinking bowls (see Gross, 1955, 1957 and
Gross and Holland, 1 960 for details). These etchings were
then used to quantify drinking bouts. Precautions were
taken in his study to reduce recording of behavior other
than drinking, but because these recordings were unat-
tended it is unclear whether all the behavior recorded was
actually drinking. Our study had the distinct advantage
of directly viewing all of the animals' behavior, thereby
allowing differentiation of drinking and exploratory be-
haviors, which enabled precise quantification of drinking
behavior.
Birgn.*; Intro usually inhabits sand burrows or piles of
decaying vegetation during the day and forages at night
when ambient temperatures are cooler (Gross, 1964).
These crabs employ both physiological and behavioral
means of osmoregulation, although the main method ap-
pears to be behavioral avoidance of desiccation (Gross,
1 964) along with uptake of water by drinking from inland
pools. Birgus lalro also uses a suite of physiological adap-
tions for osmoregulation. B. latro can reabsorb salts from
the urine (Harris and Kormanik, 1981; Greenaway and
Morris, 1 989) and it has evolved the ability to excrete uric
acid and therefore waste less water in nitrogen elimination
(Bliss and Mantel, 1968: Kormanik and Harris, 1981;
Greenaway and Morris, 1989). During periods of dehy-
dration, specimens of B. latro continue to produce isos-
motic urine and maintain intracellular fluid volume while
sacrificing extracellular stores (Burggren and McMahon,
1981; Harris and Kormanik, 1981). The large abdomen
is the water storage site in B. latro (Harris and Kormanik,
198 1 ) and becomes quite distended when fully hydrated,
but only a small volume of water is stored in the branchial
cavity relative to other terrestrial crabs (Wood and Bou-
tilier, 1985). In addition, the gills are highly reduced
(Cameron, 1981), thus limiting evaporative water loss,
and they are used as exchange sites for ions, water, and
carbon dioxide with oxygen uptake taking place at the
primitive lung (Greenaway el ai, 1988). Thus, it seems
that B. Intro's ability to differentiate between water of dif-
ferent salinities and its precise regulation of hemolymph
concentration through piecemeal drinking augment its
suite of other behavioral and physiological mechanisms
and help to explain its high degree of terrestriality.
Birgu\ lalro has never been observed drinking seawater
directly from the ocean, although tracks have been found
on dunes close to the shoreline (Gross, 1964), and Grubb
( 1971 ) reported anecdotal evidence of coconut crabs vis-
iting the ocean. Considering the evidence presented in
this study as well as in Gross (1955). B. Intro might use
the ocean as a water source under certain conditions. Field
studies investigating the natural behavior of these crabs,
particularly on some of the dry Pacific atolls, would help
bring the findings of this study into context with the nat-
ural strategies these animals employ to osmoregulate be-
haviorally.
Gecnrcoidea lalandii usually inhabits dry inland bur-
rows (Bliss, 1968) and probably relies on intermittent ac-
cess to dew and rain, along with soil water, for water up-
BEHAVIORAL OSMOREGULATION IN CRABS
1450 ,
1350
A HEMOCONCENTATED
1250
n f
1150 -
NORMAL
n n
1050 -
\ n \
-
950 -
7,
-I-
\ \ 3 \
; \
850 -
l~0
\
\ \ \ \ \
''/ \ \ \ ''' \
750 -
/''//'
\ ''/ ' ''' ''/ ''>
650 -
\ \ \ \
**'*'', *t
550 -
\ \ \ \
\ \ \ \ \
450 -
\ \ \ \
'''' '''''
350 -
250 -
150 -
SO -1
421
GZ GP G10 Gl G2 G5 G6 G7 G8 G2 GB Gl Gl G3 G4 G4 G4 G5 G7 G8 G10
350 -,
B
325 -
300 -
NORMAL HE.
u
"JPONPFNTATFT
275 -
nd\
IV
_/V-/V_^INv_/[_l v I rA 1 l_l_
250 -
= 225"n
$ 200 -
0
I 175
rP
CD 150
rf
LU
5 125 -
n
rra
r-ra
B
rra
n ^
100 •
- -
50 1
3
25 ,j
i 1
^
n .-
2 ^
-ii i yfi i
„ - .
j
^
GZ GPG10 Gl G2 G5 G6 G7 G8 G2
GB Gl Gl G3 G4 G4 G4 G5 G7 G8 G10
Figure 3. (A) Hemolymph concentrations (mOsm), and (B) weight (g), before (open bars) and after (shaded
bars) 12-hour tests involving Gecarcoidea lalandii. Left-hand regions of figure are tests with initially normal
hemolymph concentrations (<I050 mOsm) and right-hand regions are tests with initially concentrated
hemolymph osmolanties (> 1050 mOsm). X-axis labels identify individual specimens.
take (Wolcott, 1988). Although it is a highly terrestrial
crab, its gills are not as reduced as those of B. latro (Cam-
eron, 198 1 ) and are therefore more subject to evaporative
water loss. In this study, specimens of G'. lalandii preferred
to immerse themselves in. and drink from, freshwater re-
gardless of initial blood condition. This contrasts with the
water uptake strategy employed by their cogener G- na-
talis, which prefers to drink (Gibson-Hill. 1947). This is
the first quantification of drinking preference for this spe-
cies. The preference for freshwater is not surprising con-
sidering the recent work of Wolcott and Wolcott (1985.
1991) who showed that other brachyurans (Gecarcinns
lateralis and Ocypode quadrata) can reabsorb salts through
urine reprocessing in the branchial chamber. Perhaps this
and other physiological and behavioral osmoregulatory
strategies explain why brachyurans do not need to rely
heavily on seawater for their water budget. Wolcott and
Wolcott (1988) conclude that G. lateralis inhabiting the
island of Bermuda seldom if ever comes in contact with
seawater except when spawning. Further, they conclude
422
C. A. COMBS ET AL.
HI
F
55
c
UJ
g
IL
o
g
O 40
no rmal
*
hem< 'Concentrated
*
fw sw fw sw
nor nal
hemo
concentrated
fw sw
fw sw
O
i
c
~r
i
2 80
g
UJ
3 60
noi
mal
^
herao
concentrated
^ 40.
U.
O
MEAN %
? 8
r^
fw
sw
fw
sw
Figure 4. Drinking preference for freshwater or seawater according to
initial hemolymph concentration (hydrated and dehydrated) dunng 12-
hour tests for Gecarcoidea lalandii (n = 5). Drinking preference is ex-
pressed as (A) mean percentage of total immersion time, (B) mean per-
centage of total number of cheliped drinks. (C) mean percentage of total
time spent cheliped drinking.
that most of the drinking water available to land crabs on
Bermuda is freshwater, even near the edge of the shore.
Brachyurans living on some islands in the Indo-Pacific
region however, face seasonal paucity of freshwater (Gross
1964), but their behavior during these times has yet to be
reported.
Land crabs in general are euryhaline (Mantel and
Farmer, 1983) and can withstand a wide range of he-
molymph concentrations. Restoration of hemolymph os-
molarities to normal levels after dehydration occurs
quickly in both species but by different methods. This
phenomenon may be a result of the natural unavailability
of water sources (other than the ocean) at certain times
of year thus forcing animals to rehydrate quickly after
desiccation whenever favorable situations occur. This is
evidenced by the amazing amount of time (5 h av.) and
activity (>2500 individual cheliped drinking cycles) ex-
pended by B. latro in rehydrating itself after desiccation.
In contrast, B. latro spent an average of only 1 h per night
drinking, performing only about 500 individual cheliped
drinking cycles when hydrated.
G. lalandii preferred to immerse itself rather than drink
in order to obtain water, although it did occasionally drink.
The reason for this difference in strategy is unclear, but
it could have both a physiological and ecological basis. B.
latro experienced both hemoconcentration and dehydra-
tion (i.e., weight loss) when deprived of water. This sug-
gests that some tissue water loss occurs, most probably
from the large abdomen, in addition to a reduction in
hemolymph volume. Therefore, it is possible that B. latro
drinks to replenish tissue water content. G. lalandii. how-
ever, appears only to undergo hemoconcentration, and
this may be related to the fact that this species lacks an
obvious store of tissue water. As a consequence, hemo-
lymph volume may be sacrificed to maintain tissue water
content, and the major function of immersion may be to
alleviate a hemolymph ion load, which can be done readily
across the gills. Thus, the different strategies of rehydration
may be in answer to two different physiological stresses.
The decrease in time devoted to obtaining water via
immersion versus cheliped sweeps (2.5 min average fol-
lowing desiccation and 1.3 min average when hydrated
per 12-h test session) may also decrease the risk of pre-
dation in G. lalandii, which is smaller and potentially
more vulnerable than B. latro.
Further research involving behavioral osmoregulation
may investigate the feedback mechanisms of internal and
external osmoreceptors. Internal receptors that monitor
blood osmolality have not yet been found. However, in
other species, hormones produced by cells in the optic
ganglia, brain, and thoracic ganglia affect the movement
of salts and water across the gills, renal organ membranes,
and gastrointestinal wall (Hill and Wyse, 1989). Green-
away (1988) speculated that B. latro has osmoreceptors
on the chelae or mouth parts. It is quite possible that B.
latro and G. lalandii also have internal blood osmotic
receptors and that feedback mechanisms involving hor-
mones and external osmoreceptors enable it to choose
water of the appropriate salinity to maintain normal blood
osmolalities. Future work involving manipulation of
blood osmolality and testing with the video protocol we
have established might initiate understanding of the
BEHAVIORAL OSMOREGULATION IN CRABS
423
mechanisms that enable B. lalro and G. lalandii to os-
moregulate behaviorally.
Acknowledgments
We gratefully thank Lynn M. Robison. Dr. James West,
Dr. Robert Lishak, and John Newman for their contri-
butions to this investigation and to Dr. Lawrence Wit for
contributing laboratory space and for advice and criticism.
Supported by NSF DCB 88-01926 to RPH, by funds from
the NSF Research Experience for Undergraduates (REU)
program, and by the Alabama Agricultural Experiment
Station (AAES 15-923248).
Literature Cited
Bliss, D. E. 1968. Transition from water to land in decapod crusta-
ceans. Am. T-ool 8: 355-392.
Bliss, D. E. 1979. From sea to tree: saga of a land crab. Am Zool. 19:
385-410.
Bliss, D. E., and L. H. Mantel. 1968. Adaptations of crustaceans to
land: a summary and analysis of new findings. Am. Zool. 8: 673-
685.
Burggren, VV. \V., and B. R. McMahon. 1981. Hemolymph oxygen
transport, acid-base status, and hydromineral regulation during de-
hydration in three terrestrial crabs, Cardisoma. Birgus, and Coenobita.
J Exp. Zool. 218: 53-64.
Burggren, \V. W., and B. R. McMahon. 1988. Biology Of The Land
Crabs. Cambridge University Press. Cambridge.
Cameron, J. N. 1981. Brief introduction to the land crabs of the Palau
Islands, stages in the transition to air breathing. J. Exp. Zool. 218:
1-5.
Gibson-Hill. C. A. 1947. Field notes on the terrestrial crabs. Bull Raffles
Mus. 18: 43-52.
Gilles, R., and A. Pequeux. 1981. Cell volume regulation in crustaceans:
relationship between mechanisms for controlling the osmolanty of
extracellular and intracellular fluids. J. Exp. Zool. 215: 351-362.
Greenaway, P. 1988. Ion and water balance. Pp. 21 1-248 in Biology
uf the Lund Crabs. W. W. Burggren, and B. R. McMahon, eds. Cam-
bridge University Press. Cambridge.
Greenaway, P., and S. Morris. 1989. Adaptations to a terrestrial ex-
istence by the robber crab, Birgus lalro L. III. Nitrogenous excretion.
J. Exp. Biol. 143: 333-346.
Greenaway, P., S. Morris, and B. R. McMahon. 1988. Adaptations
to a terrestrial existence by the robber crab Birgus lalro. II. In vivo
respiratory gas exchange and transport. J. Exp. Bio/. 140: 493-509.
Gross, W. J. 1955. Aspects of osmotic regulation in crabs showing the
terrestrial habitat. Am. ,\'al. 89: 205-222.
Gross. W. J. 1957. A behavioral mechanism for osmotic regulation
in a semi-terrestnal crab. Biol. Bull 113: 268-274.
Gross, VV. J. 1964. Water balance in anomuran land crabs on a dry
atoll. Biol. Bull. 126: 54-68.
Gross, W. J., and P. V. Holland. 1960. Water and ionic regulation in
a terrestrial hermit crab. Physiol. Zool. 33: 21-28.
Gross, W. J., R. C. Lasiewski, M. Dennis, and P. Rudy. Jr. 1966. Salt
and water balance in selected crabs of Madagascar. Camp. Biocliem.
Physiol 17:641-660.
Grubb. P. 1 97 1 . Ecology of terrestrial decapod crustaceans of Aldabra.
Phil. Trans Roy Soc. Loud 260: 411-416.
Harris, R. R. 1977. Urine production rate and water balance in the
terrestrial crabs Gecarcinus laleralis and Cardisoma guanhumi .1
Exp. Biol. 68: 57-67.
Harris, R. R., and G. A. Kormanik. 1981. Salt and water balance and
antennal gland function in three Pacific species of terrestrial crab
(Gecarcoidea lalamiii, Cardisoma carnifex. Birgus lalro}. II. The ef-
fects of desiccation. ./ Exp. Zool. 218: 107-1 16.
Henry, R. P., and J. N. Cameron. 1981. A survey of blood and tissue
nitrogen compounds in terrestrial decapods of Palau. / Exp. Zool.
218: 83-88.
Hill, R. W., and G. A. Wyse. 1989. Animal Physiology. 2nd. ed. Harper
and Row. New York. 656 pp.
Huggins, A. K., and K. A. Munday. 1968. Crustacean metabolism.
Adv. Comp. Biochem. Physiol. 3: 271-378.
Kormanik, G. A., and R. R. Harris. 1981. Salt and water balance and
antennal gland function in three Pacific species of terrestrial crab
(Gecarcoidea lalandii, Cardisoma carnifex, Birgus lalro). I. Urine
production and salt exchanges in hydrated crabs. J. Exp. Zool. 218:
97-109.
Lister, J. J. 1888. On the natural history of Christmas island, in the
Indian Ocean. Proc. Zool. Soc. Land. 2: 512-531.
Mantel, L. H., and L. L. Farmer. 1983. Osmotic and ionic regulation.
Pp. 53-161 in The Biology of Crustacea. Vol. 5., L. H. Mantel and
D. E. Bliss, eds. Academic Press, New York.
Mangum, C. P. 1981. 1983. Oxygen transport in the blood. Pp. 373-
362 in The Biology of Crustacea. Vol. 5. L. H. Mantel and D. E.
Bliss, eds. Academic Press, New York.
Morris, S., P. Greenaway, and B. R. McMahon. 1988. Adaptations
to a terrestrial existence by the robber crab Birgus lalro. I. An in vitro
investigation of blood gas transport. / Exp. Biol. 140: 477-491.
Powers, L. W., and Bliss, D. E. 1983. Terrestrial adaptions. Pp. 271-
334 in The Biology of Crustacea. Vol. 8, F. J. Vernberg and W. B.
Vernberg. eds. Academic Press. New York.
Schoffeniels. E. 1976. Adaptations with respect to salinity. Biochem.
Soc. Symp. 41: 179-204.
Taylor, E. VV. 1982. Control and co-ordination of ventilation and cir-
culation in crustaceans: responses to hypoxia and exercise. J. Exp.
Biol. Ill: 103-121.
\\ heatly, M. G., VV. VV . Burggren, and B. R. McMahon. 1984. The
effects of temperature and water availability on ion and acid-base
balance in hemolymph of the land hermit crab Coenobita clypeatiis.
Biol. Bull. 166: 427-445.
Wolcott, D. L., 1991. Nitrogen excretion is enhanced during urine
recycling in two species of terrestrial crab. J. Exp. Zool. 259: 181-
187.
VV'olcott, T. G., 1988. Ecology. Pp. 55-95 in Biology of the Land Crabs.
W. W. Burggren. and B. R. McMahon. eds. Cambridge University
Press. Cambridge.
Wolcott, T. G., and D. L. Wolcott. 1985. Extrarenal modification of
urine for ion conservation in ghost crabs. Ocypode quadrata (Fabri-
cus). / Exp. Mar. Biol. Ecol. 91: 93-107.
Wolcott, T. G., and D. L. Wolcott. 1988. Availability of salts is not a
limiting factor for the land crab Gecarcinus Laleralis (Freminville).
J. Exp. Mar. Biol. Ecol. 120: 199-219.
Wolcott, T. G., and D. L. Wolcott. 1991. Ion conservation by repro-
cessing of urine in the land crab Gecarcinus lateralis (Freminville).
Physiol. Zool. 64: 344-361.
Wood, C. M., and R. G. Boutilier. 1985. Osmoregulation, ionic ex-
change, blood chemistry, and nitrogenous waste excretion in the land
crab Cardisoma carnifex: a field and laboratory study. Biol. Bull
169: 267-290.
Wood, C. M., R. G. Boutilier, and D. J. Randall. 1986. The physiology
of dehydration stress in the land crab, Cardisoma carnifex: respiration,
ionoregulation, acid-base balance and nitrogenous waste excretion.
J.Exp. Bwl 126:271-296.
Yancey, P. H., M. E. Clark, S. C. Hand, R. D. Bowlus, and G. N. Somero.
1982. Living with water stress: evolution of osmolyte systems. Sci-
ence 21T. 1214-1222.
Reference: Biol. Bull. 182: 424-434. (June. 1992)
Causes and Consequences of Fluctuating
Coelomic Pressure in Sea Urchins
OLAF ELLERS1 AND MALCOLM TELFORD2
Department of Zoology, University of California. Davis. California 95616 and
2 Department of Zoology, University of Toronto, Toronto. Ontario M5S 1A1, Canada
Abstract. We measured coelomic pressure in sea urchins
to determine whether it was high enough to support a
pneu hypothesis of growth. In Strongylocentrotits purpur-
atus the pressure was found to fluctuate rhythmically
about a mean of -8 Pa, and was negative for 70% of the
time. This is at variance with the theoretically required
positive pressures of the pneu hypothesis. Furthermore,
there were no sustained significant differences between
the pressure patterns of fed and starved urchins, presumed
to be growing and not growing, respectively. The rhyth-
mical fluctuations in pressure were caused by movements
of the lantern which changed the curvature and tension
of the peristomial membrane. We developed a mathe-
matical and morphological model relating lantern move-
ments, membrane tension, and pressure, that correctly
predicts the magnitude of the fluctuations. Pressures pre-
dicted by the model depend also on coelomic volume
changes. In Lytec/iiniis variegatitx simultaneous retraction
of the podia, which causes expansion of the ampullae,
resulted in an 8.8 Pa increase in coelomic pressure, relative
to the pressure during simultaneous podial protraction.
Introduction
For some seventy-five years, the growth and shape of
sea urchins have, with few exceptions, been attributed to
a similarity with internally pressurized tensile structures.
D'Arcy Thompson (1917) remarked on the similarity of
shape between sea urchins and water droplets on a glass
plate. A water-filled balloon resting on a table (Fig. 1 )
provides an analogous form. This basic idea has been in-
voked repeatedly to explain both growth and form. Moss
and Meehan (1968) suggested that growth of the gut and
gonads increased coelomic pressure and this caused
Received 21 October 1991; accepted 27 March 1992.
growth in the test. Likening echinoids to inflated structures
(pneus), Seilacher (1979) argued that variations in shape
among regular and irregular echinoids could be explained
by forces from the tube feet and by the occurrence of
internal "tethers" of calcite or collagen. Dafni and Erez
(1982), Dafni (1983. 1985, 1986). and Baron (1988), all
assumed the existence of positive internal pressure in sea
urchins, and explained morphogenesis in terms of the re-
sulting stress patterns and the action offerees from other
sources such as podia, internal muscles, and mesenteries.
Although internal fluid pressure is usually not relevant
in the functional analysis of solid structures, there are
engineering designs in which it does play an important
role. While designing underwater storage vessels that re-
quire a minimum of wall materials, Royles el at. (1980)
were impressed by the similarity of their theoretically de-
rived shapes and some sea urchins (most notably Echinus
esculent us). The design of such "constant strength" or
"buckle-free" structures involves balancing pressure dif-
ferences (positive or negative) across the vessel wall with
forces in the wall. It is tempting to interpret the conver-
gence on an echinoid form as indicative of an underlying
similarity in the balance offerees. Royles et al. (1980)
actually coined the expression of "Echinodome" for these
structures.
The obvious and crucial question — what is the mag-
nitude of the internal pressure in echinoids — has not been
answered. Dafni (1985, 1986) attempted to manipulate
forces acting on the growing test and isolated plates, but
provided no measurements of pressure. Reporting the only
pressure measurements. Baron ( 1 99 1 ) recorded fluctuating
coelomic pressures in an echinoid. With the aid of a finite
element method he developed a complicated tensile
growth model which, although elegantly refined, is still
fundamentally a pneu hypothesis. According to his model.
424
COELOMIC PRESSURE IN URCHINS
425
Figure 1. (a) A balloon filled with water in water; (b) a balloon filled
with water in air; and (c) an urchin test. Note the similarity of shape
between the (b) and (c). The difference in shape between (a) and (b)
illustrates the importance of self-weight forces. There are no self-weight
forces on a water-filled balloon in water since the water inside and outside
are equally dense. In urchins, the internal volume also has no effective
weight; thus the downward forces result only from the underwater weight
of the calcite or the pull of tube feet. The weight forces are balanced by
internal pressure resulting from tension in the membrane. None of these
structures are pneus because they are not air-filled, but (a) and (b) cer-
tainly, and (c) possibly, form their shape as a result offerees analogous
to those in a pneu. including internal pressurization.
growth can occur only during periods of positive internal
pressure.
In this paper we describe a technique for measuring
coelomic pressure in sea urchins and report the results of
two series of experiments. The first series was undertaken
to determine whether there was sufficient positive pressure
to support the pneu hypothesis of growth. For this, we
compared pressures in sea urchins (Stronglyocentrotus
pitrpiiratus) fed ad libitum and presumed to be actively
growing, with pressures in starved animals, presumably
not growing (Ebert, 1968). After measuring the fluctuating
pressures, we investigated the possible morphological and
physical causes of the pressure patterns. This led to de-
velopment of a model relating pressure changes to alter-
ations in curvature in the peristomial membrane during
protraction and retraction of the lantern. In the second
series of experiments we examined the effect of volume
changes, resulting from the alternate extension and re-
traction of podia, on coelomic pressure in Lytecliinux vur-
iegatus. We consider the interaction of volume changes
and behavior of the peristomial membrane in explaining
the observed pattern of coelomic pressures in sea urchins.
Materials and Methods
Experimental animals
Specimens of Strongylocentrotus purpuratus collected
subtidally at Bodega Bay, California, and maintained in
running seawater, were divided into two lots. The first
was fed ad lihitiim with kelp (Macrocystis sp.) and the
second was starved. There were no significant differences
in the size of urchins in the fed (33.0-81.4 mm. n = 27)
and unfed (41.9-82.6 mm. n = 25) groups. Size was es-
timated by a volume approximation which was (height
X diameter)2. Pressure measurements were performed
during a three-week period, starting two months after the
beginning of these feeding regimes. Lytechinus variegatus
(53.9-68.1 mm diameter) was collected at Long Key,
Florida, and maintained on natural substrate with dead
leaves of Thalassia tcstmlimim, for 12 to 72 h before ex-
perimental use.
Pressure measurement
Internal pressure was measured by mounting the ur-
chins on a vertical. 14 gage, hypodermic needle passing
through the peristomial membrane. The needle was con-
nected to one side of a P305D differential, moving mem-
brane, pressure transducer (Validyne Corporation.
Northridge. California) fitted with a nickel plated 3-20
membrane to read pressures up to ±550 Pa. The other
side of the transducer was open to the seawater surround-
ing the experimental animal.
Calibration of pressure transducer
The system was calibrated before each series of mea-
surements. Calibrations and all experiments were per-
formed in a two-chambered Plexiglas aquarium. At the
start, seawater levels in the two chambers were equilibrated
via a connecting valve. After closure of the valve, the water
426
O. ELLERS AND M. TELFORD
level in one chamber (positive side of transducer) was
raised by increments of 1 . 1 mm by the gradual immersion
of a Plexiglas box propelled by a threaded drive mecha-
nism. At each step the voltage output at 1-s intervals was
averaged over a 30-s period by a Dynamic Signal Analyzer
(Hewlett-Packard #3561 A). Initial calibrations were con-
tinued to a total pressure head of about 22 mm of seawater
(220 Pa). Later calibrations extended only to 1 1 mm of
seawater, which adequately covered the range of pressures
commonly encountered. Calibration readings were taken
as pressure increased and as it decreased back to zero.
Linear regression of transducer output (mv) and pressure,
fitted by least squares, was used to convert experimental
readings to pressure. For field experiments in Florida, the
system was simplified. The Plexiglas box and threaded
drive assembly was replaced by a pipetting technique in
which 1 5-ml aliquots of seawater were added sequentially
and then removed from the reference chamber.
Estimate of errors in pressure measurements
Due to uncertainty in the measurement of the pressure
head against which the transducer was calibrated, the
range of bias in the slope of the calibration curve was less
than 0.1%. The precision range of the slope was ±10%
because of day-to-day variation. Additionally, in the worst
case, the 8-bit digitizer recorded only to the nearest 1.7
Pa, and there was drift in the zero; a combined imprecision
range of ±3 Pa resulted. The accuracy can be expressed
as ±(10.1% + 3) Pa.
We were concerned that urchins might leak, thus ar-
tificially relieving high positive or negative pressures. We
ruled out this possibility by injecting the urchins with food
coloring and by coloring the liquid in the transducer. We
observed no color leakage, except at very much higher
pressures than those reported in this experiment.
Internal pressure could also be artificially relieved by
flow through the needle into the tiny space vacated as the
metal membrane of the transducer shifted while making
the measurement. This possibility was minimized by use
of a "low volume" pressure transducer. To test this po-
tential error, we set up an experiment in which we could
simulate the pressure measurement and watch what hap-
pened to the pressure and volume. The urchin was re-
placed by a rubber tube filled with dyed seawater, closed
at one end, and attached to a 5 mm diameter graduated
pipet that was open to the atmosphere at the other end.
With fluid in the pipet levelled to measure 40 Pa, we in-
serted the needle through the rubber hose. There was no
detectable motion of the water level in the pipet, indicating
that volume changes due to the transducer motion were
less than 3 jil; in a 60 mm diameter urchin, this volume
change could be accommodated by a 10 ^m upward or
downward motion of the lantern involving a strain of
5 X 10 6 in the peristomial membrane, an amount that
has a negligible effect on pressure in the coelom.
Experimental procedure
Each urchin, when mounted on the needle, rested on
a small platform. The podia reached the platform but
could not reach the sides or the floor of the aquarium.
During the course of an experiment the transducer output
was sampled at 5. 12 Hz and digitized. The trace was dis-
played by the signal analyzer simultaneously with a fre-
quency spectrum. The data were transferred in 200-s sec-
tions to an Apple Mac II equipped with a "LabVIEW"
GPIB interface card (National Instruments, Austin,
Texas). For each urchin, data were recorded for 10 min.
The zero point of the transducer was checked after each
measurement was completed, and the needle was detached
and syringed to remove any coagulated coelomic fluid.
Diameter and height of each specimen was measured by
calipers. The water in the experimental chamber was re-
placed after each group of five specimens to minimize
changes in water temperature.
The procedure for L. variegatus was similar except that
a 10 min section of data was transferred directly into the
computer, and the light level was manipulated to induce
podial movements. For each often urchins, room lights
and fiber-optic microscope lights directed at the urchin
were alternately switched on and off every 2 min. When
the lights were on, the podia retracted; when the lights
were off, they extended.
Data analysis
For S. piirpuratus specimens, each 200 s trace was
scanned and the following information was compiled: (i)
seconds below zero pressure; (ii) the mean pressure; (iii)
the standard deviation of pressure; (iv) the maximum
pressure; (v) the minimum pressure; (vi) the mean of pos-
itive pressures; (vii) the standard deviation of positive
pressures; (viii) the mean of negative pressures; (ix) the
standard deviation of negative pressures. Two-way anal-
yses of variance by trace and by feeding regime were per-
formed on these data. Additional /-tests were performed
to compare fed and starved animals by successive traces.
A Fourier transform of the third 200-s trace for each spec-
imen gave the amplitude and periodicity of rhythmic
pressure fluctuations. Using the first 200-s trace (during
which the needle was inserted), a discriminant functions
analysis was performed to see whether fed and unfed in-
dividuals could be identified from their initial pressure
patterns. We performed a stepwise regression to determine
which variables to include in the discriminant functions
analysis. The discriminant model is
Y = b
• 39X9,
(1)
COELOM1C PRESSURE IN URCHINS
427
unfed
(a)
(b)
400
450
500
250
550
fed
600
(c)
(d)
200
-80
400
600
Figure 2. (a) Pressure-time trace for an unfed urchin during the first
200 s of the experiment. The large negative pressure pulse, characteristic
of unfed urchins, occurred just after the needle was inserted through the
peristomial membrane, (b) Pressure-time trace for an unfed urchin 400-
600 s after the start of the experiment. This trace shows the characteristic,
rhythmic fluctuations of pressure associated with movements of the lan-
tern, (c) Pressure-time trace for a fed urchin during the first 200 s of the
experiment, showing the characteristic, positive pressure pulse as the
needle was inserted through the peristomial membrane, (d) Pressure-
time trace for a fed urchin 400-600 s after the start of the experiment,
showing rhythmical changes with lantern movements. Differences in the
traces for starved and fed urchins (a and c) were statistically significant;
during the third 200-s traces (b and d) the differences were not significant.
where y is equal to - 1 if an urchin is fed, and is equal to
+ 1 if an urchin is unfed. The nine variables descriptive
of the pressure traces are x, to x9. The fitted slopes are a,
to a9 and b is the intercept.
For L. variegatus the average level of pressure was
measured for each 2-min segment except the first, which
was assumed to be a settling-down period. A paired /-test
was done on the average pressures to compare the lights-
off periods with the immediately ensuing lights-on periods.
Results
Description of the pressure traces
Pressure traces for S. purpiiratus characteristically fluc-
tuated at a frequency of 0.055 Hz with a S.D. of 0.021
Hz (n = 167 traces). This corresponds to an average period
of 18 s, and the range of periods corresponding to the
above S.D. is 13-29 s.
When the needle was inserted through the peristomial
membrane, there was usually a negative or positive pres-
sure peak (Fig. 2) that often went off-scale on the recording
equipment, and that differed significantly from the fluc-
tuations in the second and third traces as shown by the
maxima and minima in Table I. Over several minutes
the pressure tended toward, and eventually stabilized at,
an average mean pressure of -8.2 Pa with a S.D. of 1 1
Pa (n = 52 urchins). According to our error estimate, zero
lies in the range ± (8.2 X 10% + 3) Pa: a /-test shows that
the worst-case zero of -3.8 Pa is significantly different
from -8.2 Pa with a S.E. of 1 .4 Pa (P < 0.01 ). The average
S.D. of the pressure was 10 Pa with S.D. of 6.4 Pa (n
= 52). The pressure was below zero 70% of the time.
Urchins fed ad libitum, and those receiving food only
via occasional cannibalism, had very different initial pres-
sure responses (Fig. 2). Well-fed urchins had pressures
that tended to increase initially. Unfed urchins had pres-
sures that tended to decrease initially. All of the variables
except S.D. differed significantly in the first 200-s trace
(Table I). Step-wise regression of variables for the first
trace indicated that the mean of the positive pressures
and the minimum pressure (r = 0.41, slope significantly
non-zero, P < 0.001 ) correctly predicted whether the an-
imals were fed or unfed 83% of the time.
There were no significant correlations between urchin
volume and any of the nine descriptive variables in any
traces for fed urchins, nor in the first 200-s trace for unfed
urchins. However, in subsequent traces from unfed ur-
chins, five of the variables (mean, S.D., minimum, mean
negative, and S.D. of negative pressures) were correlated
with test size (Table II).
Podia! movements and pressure
When the lights were turned off, L. variegatus pro-
tracted its podia and the coleomic pressure decreased.
When the lights were turned on, podia retracted and the
coelomic pressure increased (Fig. 3). Coelomic pressure
Table I
Resii/ls oft-lesls showing statistically significant differences between
fed and starved Strongylocentrotus purpuratus /or the nine variables
descriptive of coelomic pressure during the three successive 200 s traces
Variable
Trace 1
Trace 2
Trace 3
Seconds below zero
**
n.s.
n.s.
Mean pressure
***
n.s.
n.s.
S.D. pressure
n.s.
n.s.
n.s.
Maximum pressure
##
n.s.
n.s.
Minimum pressure
***
n.s.
n.s.
Mean +ve pressure
#**
n.s.
n.s.
S.D. +ve pressure
***
n.s.
n.s.
Mean -ve pressure
**
n.s.
n.s.
S.D. -ve pressure
**
n.s.
n.s.
(n.s. not significant; **/> < 0.01; ***/> < 0.001).
428
O. ELLERS AND M. TELFORD
Table II
Correlation coefficients between body si:e and statistical variables
descriptive of pressure traces from unfed Strongylocentrotus
purpuratus
Variable
Trace 1
Trace 2
Trace 3
Seconds below zero
n.s.
n.s.
_
n.s.
Mean pressure
n.s.
-0.4
-0.5
*
S.D. pressure
n.s.
0.4 *
0.4
*
Maximum pressure
n.s.
n.s.
—
n.s.
Minimum pressure
n.s.
-0.4 *
-0.5
**
Mean +ve pressure
n.s.
n.s.
—
n.s.
S.D. +ve pressure
n.s.
n.s.
—
n.s.
Mean — ve pressure
n.s.
-0.5 *
-0.6
**
S.D. -ve pressure
n.s.
0.4
0.4
*
(n.s. not significant; *P < 0.05; **P < 0.01). Note: There were no
correlations between any of these variables and body size in fed urchins.
during the lights-on period was 8.8 Pa higher than the
mean pressure during the immediately subsequent lights-
off period (P < 0.0001; n = 20: 10 urchins, 2 paired sam-
ples each).
Discussion
The fluctuating coelomic pressures observed in this
study were predominantly negative. In the wide range of
animals surveyed by Trueman (1975). most reported
pressures are positive, the highest being 104 Pa in the lug-
worm, Arenicola marina. In soft-walled pressure vessels,
the internal pressure can only be zero or positive relative
to the outside. At zero relative pressure, the body wall is
limp and any process tending to a negative internal pres-
sure will cause the membrane to collapse and fold, thus
reducing the pressure to zero (Clark and Cowey, 1958).
Negative pressures are possible in systems in which the
walls have flexural stiffness, as is the case with some skel-
etal and muscular tissues. Trueman (1975) reported pres-
sures of -500 Pa from underneath the foot of Pate/la sp.
during the passage of pedal waves. Negative pressures have
also been generated inside the gastropod foot (Voltzow,
1986) and by the suckers of an octopus (Kier and Smith,
1990; Smith. 1991). Many soft-bodied animals have some
hard, stiff parts, while many primarily hard-bodied or-
ganisms have some soft, flexible membranes. Sea urchins,
having a hard test and large peristomial membrane, are
examples of the latter.
There are several processes that could influence coe-
lomic pressures in sea urchins, but some of them do not
produce pressures of the observed magnitude. However,
we found two processes of great importance: the exertion
offeree on the coelomic fluid (for instance, by the peri-
stomial membrane) and the movement of water into the
coelomic space (as in the simultaneous retraction of the
podia). Before considering these two in more detail, we
show why a number of the other possibilities are not sig-
nificant.
Causes of pressure in urchins
Pressure is a force magnitude per area. In non-accel-
erating fluids, at each point in the fluid there is a balance
offerees in all directions. Gravitational pressure, p,,, at a
given depth is
P., = Pgd, (2)
where p is the density of seawater, g the acceleration due
to gravity, and d the depth (atmospheric pressure is not
included). We measured the difference between pressures
inside and outside the urchin. Because the two locations
were at the same depth, hydrostatic, gravitational pressures
are irrelevant, and the remaining discussion refers only
to relative transmural pressures.
Sound or sudden impacts from waves could also cause
internal pressure. The rhythmic, 20-s pressure patterns
we observed cannot be sound because there was no such
rhythm when the needle was removed from the urchin.
Nevertheless, in the ocean, sudden coelomic pressures
from impact forces such as waves and sound are possible
and might have implications for behavior, mechanical
functioning, or even pressure-regulated growth of urchins.
These phenomena have not been investigated.
Hydrodynamic forces are unlikely to be of importance
in explaining pressures inside urchins, because rates of
flow are very slow. Hanson and Gust (1986) measured
rhythmic flows inside urchin coeloms that have the same
periodicity (20 s) as the pressure pulses we measured.
Thus, fluid dynamic pressures cannot be immediately
ruled out in explaining the observed pressure patterns.
Expected pressures from flow are less than or equal to the
dynamic pressure, p,/. which is
P -
PJ = T u-
(3)
where p is the density of seawater, and u is the velocity
of flow (Vogel, 198 1 ). In our experimental observations.
i — — r
200 300 400
TIME (seconds)
600
Figure 3. The pressure pattern in Lytechinus varnyaliis when lights
are alternately turned on and off at 2-min intervals (black bar indicates
lights on). The podia protracted when the light was off and retracted
when the light was turned on.
COELOMIC PRESSURE IN URCHINS
429
the standard deviation of pressure was 10 Pa. This would
correspond to a minimum flow of 100 mm s '. Because
Hanson and Gust (1986) observed a maximum flow of
1.5 mm s '. we conclude that the pressures we observed
were not due to flow.
Tension in a curved, stretched membrane can be an-
other cause of pressure differentials. According to La-
place's law (see Popov, 1976; Wainwright el al.. 1976;
Vogel. 1988; or Ellers and Telford, 1991). the pressure
drop across such a membrane or a flexible body wall de-
pends on its tension and radius of curvature. The pressure
inside the membrane will be positive with respect to ex-
ternal pressure when the membrane is inwardly concave.
In a cylinder the pressure difference. Ap, across the mem-
brane is
(4)
where r is the radius of curvature and T the tension in
the membrane. The tension. T, is the stress times the
thickness of the material. More generally, in a three-di-
mensional shape such as a sphere or ellipsoid, two radii
of curvature are involved, so that at every point on the
surface
s ,
Ap = - +
T:
(5)
where T, is the tangential tension in one direction with
radius of curvature r,, and T: is the tangential tension in
an orthogonal direction, with radius of curvature r:
(modified from Timoshenko and Woinowsky-Krieger.
1959. p. 435). Both negative and positive differences can
occur across a membrane, depending on whether its radii
of curvature are positive or negative.
If the several coelomic compartments in echinoids (so-
matocoels. hydrocoel. axocoel, and peripharyngeal coe-
lom) (Hyman, 1955; Smith, 1984) are bounded by
stretched membranes, there is potential for a diversity of
pressure relationships between them. We found no reason
to suspect that there are more than two functionally pres-
surized spaces, the water-vascular system and the coelom
proper. Injection of red dye confirmed a separate peri-
pharyngeal space, but the membrane is flaccid and flimsy
and could not support separate pressurization. The only
stretched membranes are found in the peristome, peri-
proct. and water vascular system.
Pressure and [H'ri\ioniia/ membrane
The peristomial membrane is a circular sheet composed
of cross-fiber collagen arrays and circular and radial mus-
cles (Hyman. 1955). In some species, it contains calcite
plates or spicules (Smith. 1984; Candia Carnevali ct a/..
1990). It is joined to the test at the distal edge, and to the
lantern centrally. Thus the shape of the membrane is like
a washer: flat, with a hole in the middle. No one has stud-
ied the deformation of this membrane as the lantern pro-
tracts and retracts, but from our pressure measurements
and the general rules about membranes given above, we
can make predictions about its curvature.
Curvature of the membrane depends on the relative
pressure difference across it. As the lantern protracts, the
pressure inside becomes negative relative to ambient.
From Laplace's law, we know that a negative internal
pressure implies that the membrane is convex on the coe-
lomic side. Conversely, a positive internal pressure would
imply that the membrane is concave on the coelomic side.
The same is true for the periproctal membrane. In species
in which the periproct is flexible, its shape might indicate
a positive or negative internal pressure. These predictions
hold only if the membranes have low flexural stiffness.
Often flexural stiffness may be conferred by catch-collagen
or ossicles. If the membranes are flexurally very stiff, then
they may produce negative or positive pressures regardless
of their curvature, just as the test does not reverse its cur-
vature as internal pressure changes from positive to neg-
ative. It should be a goal of future studies to determine
the flexural stiffness of such membranes.
Regardless of the membrane curvature and flexural
stiffness, protractor and retractor muscles controlling the
motion of the lantern exert forces that cause tension in
the peristomial membrane and thus a pressure drop across
it. We observed the lantern moving in and out during our
pressure measurements, and the 20-s pressure rhythm ap-
peared to match its protraction and retraction. Jensen
( 1985) suggests that the role of such lantern movements
is to stir the coelomic fluid, thus facilitating distribution
of nutrients and respiratory gases.
Pressure and podial movements
When many podia simultaneously retract, water pre-
viously in the podia will be stored in the ampullae, thus
effectively moving water into the coelomic space. If the
peristomial membrane and periproct do not move com-
pensatorily outward, and if there is negligible How via the
madreporite. the pressure in the coelom must rapidly in-
crease. In fact, because of the incompressibility of water,
if there is no volume regulation the urchin must either
spring a leak or the pressure would become so great that
the podia could not retract. Fechter (1965) recognized
this problem. He calculated that the volume made avail-
able when the peristomial membrane moves outward is
sufficient to compensate for the volume of water moved
into the coelomic space when all podia simultaneously
contract. Further, he showed that the size of the peristo-
mial membrane was more closely correlated with the
number of podia than with test size. Finally, he demon-
430
O. ELLERS AND M. TELFORD
strated only very small flows via the madreporite during
simultaneous podial retraction. We observed that simul-
taneous podial retraction caused an 8.8 Pa pressure in-
crease in the coelom. Fechter (1965), working with
Echinus esculentus, reported an increase of 200 Pa.
Although the madreporite is not involved in volume-
related pressure regulation. Fechter (1965) concluded that
it was involved in non-volume-related changes due to
gravitational, hydrostatic pressure. We believe that Fech-
ter's conclusion must be wrong, but first we will present
his experimental evidence. Fechter glued the madreporite
shut and performed two manipulations. ( 1 ) He increased
the hydrostatic, gravitational pressure by increasing the
depth at which the urchin was kept. When the external
pressure increased the podia collapsed. (2) He pulled the
lantern outward, decreasing the pressure in the coelom,
and again the tube feet collapsed.
In the second case, the madreporite could not relieve
the induced pressure change because, according to Fet-
ter's own results, it allows insufficient flow. We argue,
instead, that pulling the peristomial membrane outwards
causes a volume flow from the podia into the ampullae.
In the first case, when hydrostatic pressure increases, it
does so with negligible volume change. Therefore, al-
though the increase in hydrostatic pressure may be suf-
ficient to cause the podia to collapse, it would do so only
if the pressure was being relieved by a flow from the podia
into the ampullae. But because this pressure change is
gravitational, it is not associated with a volume change,
and therefore even the tiniest flow from the podia into
the ampullae will immediately relieve the pressure differ-
ence.
The only way we can explain Fechter's results is if there
was an air bubble in the coelom that would have dimin-
ished in size with increasing gravitational pressure, there-
fore causing flow from the podia into the ampullae. Such
air bubbles sometimes form in urchins that have been in
air for some time. Fechter dried the madreporite with a
stream of hot air, before gluing it shut. Perhaps this pro-
cedure explains his results. We suggest that, contrary to
Fechtefs conclusion, his experiments do not show that
the madreporite functions to accommodate hydrostatic
gravitational pressures. Furthermore, such a function is
unnecessary because volume changes caused by hydro-
static pressure would be accommodated by miniscule
flows and deformation of tissues.
Although accommodation of hydrostatic, gravitational
pressure is unnecessary, there are other types of pressure
that might require the coelomic pressure to be maintained
independent of the water- vascular system, and perhaps
the madreponte has such a role. For instance, the pressure
fluctuations we observed (±10 Pa) could have caused the
podia to malfunction because these pressures would be
exerted on the ampullae inside the coelom. But such fluc-
tuations can only cause podia to extend or retract if they
cause the ampullae to expand or contract, which would
happen only if volume changes were associated with the
pressure fluctuations. Additionally, the deformation of a
membrane depends on its stiffness and on radius of cur-
vature [as in equations (4) and (5), above]. The radius of
curvature of the ampullae is much smaller than that of
the peristomial membrane, and therefore we expect much
smaller deformations in the ampullae. That the ampullae
have a smaller radius of curvature than the peristomial
membrane may be a design requirement of echinoderm
water-vascular systems.
The digestive tract is another potential source of pres-
sure change. When full, the stomach will take up more
room in the coelom, and the peristomial membrane must
move outwards to relieve the volume increase. Similarly,
flows into and out of the mouth, or in the siphon, may
cause volume fluctuations that could cause pressure
changes if the peristomial membrane does not move
compensatorily. Further, without compensation by the
peristomial membrane, defecation may lower coelomic
pressure because it tends to reduce the volume of gut con-
tents.
Finally, several authors have described ruffled sacs
hanging externally from the peristomial membrane (Hy-
man, 1955; Smith, 1984), the supposed function of which
is either as gills or pressure regulators for the peripharyn-
geal coelom. However, no experimental data about their
function have been presented. We saw no evidence that
these sacs expanded or contracted while the coelomic
pressure fluctuated. Furthermore, their openings are far
too small to allow sufficient flow to regulate coelomic
volume.
A model of forces causing a pressure drop across the
peristomial membrane
The forces causing protraction of the lantern, and thus
tension in the peristomial membrane, come from lantern
protractor and retractor muscles and from the submerged
weight of the lantern. These forces must be estimated.
Andrietti ct al. (1990) report 3 g (0.03 N) for lantern weight
minus buoyancy in a specimen of Paracentrotus lividus.
They also report forces of 40 g (0.4 N) exerted by lantern
protractors and forces of 10 g (0.1 A7) exerted by lantern
retractor muscles. Because P. lividus rarely exceeds 70
mm diameter (Mortensen, 1977), it is similar in size to
51 purpwatus and L. variegatus. and the forces should be
comparable.
The assumed geometry of the lantern, test and peristo-
mial membrane are shown in Figure 4a. The forces on
the peristomial membrane are: ( 1 ) a vertical force, ft , ex-
erted by the lantern weight and the lantern muscles; (2)
forces from the pressure difference across the membrane;
and (3) the reactive, tensile force exerted on the membrane
COELOMIC PRESSURE IN URCHINS
431
(a)
for the present context. The two-dimensional approach
used here should give results of the correct order of mag-
nitude.
The radius of curvature of the peristomial membrane,
rpm, for a given protraction of the lantern, v, and a given
horizontal, peristomial radius, h, can be derived from the
geometry shown in Figure 4b. The radius of curvature is
(b)
Figure 4. (a) Location of the peristomial membrane in an urchin.
The star in both figures marks the point of attachment to the edge of
the penstome. (b) Geometric model of the penstomial membrane. The
angle 6, and the radius of curvature of the membrane are not independent.
Zero vertical displacement occurs when the membrane is horizontal.
by the test. The vertical force, ft,, exerts a force, f,,,, in the
membrane.
(6)
cos (6) '
where 0 is the angle between the vertical and a tangent at
the central margin of the membrane (at the point of at-
tachment of the peristomial membrane to the teeth) (Fig.
4b). The force, f,,,, on the membrane corresponds to a
tension, T, (force per length) in the membrane of
T =
(7)
where r, is the radius of the central margin of the peristo-
mial membrane. From Laplace's equation (4)
Ap = — ,
(8)
where r,,,,, is the radius of curvature of the membrane. In
using equation (4) rather than (5) we make two simplifying
assumptions: that a second horizontal radius of curvature
can be ignored, and that the curve formed by a vertical
cross section of the peristomial membrane has a single
radius of curvature at every point. In reality this curve
may have variable radii of curvature. A more realistic
model would add an unjustifiable degree of complexity
2 cos (arctan (v/h)) cos (6 + arctan (v/h))
Substituting through equations 6, 7 and 8,
Ap =
ft, cos (arctan (v/h)) cos (6 + arctan (v/h))
Trhr, cos (6)
(9)
(10)
which is shown in Figure 5. This graph shows that many
possible combinations of pressure, protraction, and 9 are
possible when only the force balance on the membrane
is considered. Initially, this may seem counterintuitive.
Intuition suggests that as the lantern protracts, the internal
pressure should get more and more negative relative to
outside as the membrane pulls more and more on the
constant volume of water inside the urchin. That this
pressure pattern is not implied in Figure 5 reflects the fact
10
£
L=1.0
71/4
membrane angle, 8
7C/2
Figure 5. Contour plot of theoretical predictions from the geometric,
force balance model of the peristomial membrane (see Fig. 4 and text
for details). Elongation of the penstomial membrane and pressure across
it are functions of the membrane angle at the central edge, 0, and pro-
traction, v, given a downward force of the teeth and lantern muscles on
the membrane, ft.. This graph shows that many combinations of 6 and
v are possible at a given pressure across the membrane. Which 0, v path
the membrane follows as the lantern protracts depends on the volume
of the urchin and the properties of the penstomial membrane.
432
O. ELLERS AND M. TELFORD
that the force balance makes no assumption about the
volume of water inside the urchin, nor about the material
properties of the peristomial membrane.
To understand a fluctuating pattern of pressure be-
coming increasingly negative as the lantern protracts, ex-
amine the change in length of the peristomial membrane.
The length of the membrane, the distance along its vertical
arc from the attachment point at the test to its attachment
point at the teeth, is
2(0
(11)
(the angle 4> is shown in Fig. 4b).
By examining the contour plots of pressure drop, and
peristomial membrane length (Fig. 5), it is possible to
imagine what is happening as the lantern moves. As it
protracts, the peristomial membrane elongates, and, as-
suming constant coelomic volume, the internal pressure
must decrease. Initially, assume that the membrane starts
at the point, ft = TT radians (the membrane is straight and
horizontal). As the lantern protracts, the line representing
the motion of the lantern must move towards higher v
(protraction) and towards lower 0 on the graph, to stay
in the region of negative pressure and simultaneously to
increase the length of the peristomial membrane. Increase
in the length of the peristomial membrane helps to com-
pensate for volume changes that would otherwise occur
because it can arch upward, effectively compensating for
the volume of the lantern pulled downward.
According to Figure 5, the tendency to decrease 0 while
increasing v, initially causes Ap to become negative
quickly because many pressure contour lines must be
crossed, but, after even a little protraction, it is possible
for the lantern to protract and follow an isobar. This may
be an explanation for the plateaus often observed at the
peaks of fluctuations in the pressure trace. A path followed
by the lantern could be specified by two functions of time,
fl(time) and v(time), which we call a "0, v" path. This
path, represented by an imaginary line in Figure 5. will
depend on the constraints imposed by the degree of con-
stancy of the coelomic volume and the material properties
of the peristomial membrane. We plan to develop this
theoretical model further in the future and obtain mea-
surements of the motion of the lantern, the constancy of
the coelomic volume, and the material properties of the
peristomial membrane.
This crude, initial model serves to explain some aspects
of the relationship between pressure and the behavior of
the structures that cause it. The pressures are of the correct
order of magnitude to have been caused by lantern mus-
cles. The mean negative pressure observed (-8 Pa) is small
enough that it could have been caused by the weight of
the lantern. If the podia simultaneously retract, or if the
stomach is full, thus raising the coelomic volume, this
model shows that the lantern can still protract with only
a change in the H, v path. Finally, it is reasonable to cal-
culate the pressure based solely on what the peristomial
membrane is doing, because the pressure inside the ur-
chin's coelom is the same everywhere, and thus if any
other structure were contributing, it would have to be
balanced by tension in the peristomial membrane.
Implication* of the observed pressures for the pneu
theory
In keeping with the pneu hypothesis, we expected con-
tinuously positive internal pressure in sea urchins. Instead
we found fluctuating positive and negative pressures, with
an overall mean below zero. Clearly, the original version
of the pneu hypothesis must be rejected on the basis of
these measurements.
Baron ( 1 99 1 ) also recognized the problem for the pneu
hypothesis when he found fluctuating pressures. He de-
veloped a modified version of the hypothesis that preserves
the spirit of the original (Thompson, 1917) but incorpo-
rates new rules for growth of the skeleton. Baron (1991)
proposed that skeletal plates grow at their margins when-
ever they are in tension, and that growth is directly pro-
portional to tensile stress. Instead of the term "pneu." he
called this a "tensile growth model." These growth rules
necessitated development of a finite element analysis to
determine the expected stresses in the skeleton caused by
internal pressure and other forces, such as those from tube
feet. From these analyses Baron ( 1991 ) was able to gen-
erate urchin-like shapes using a computer. Making several
alternative assumptions about internal pressure, he ex-
amined their effect on the shapes produced by his model.
In these simulations, he found that a pressure fluctuating
about a mean of 30 Pa, with a S.D. of 30 Pa, generated a
shape indistinguishable from that produced with a con-
stant pressure of 30 Pa. Based on this finding, he thereafter
simulated urchin shapes using constant pressures.
Baron's ( 1 99 1 ) assumptions can be compared with our
more extensive pressure measurements. For his standard
growth situation he assumed a pressure of 30 Pa. and the
other pressure used was 1 5 Pa. We observed an average
pressure of -8 Pa. Under fluctuating pressure regimes he
assumed a negative pressure for at most 17% of the time,
whereas we observed it for 70% of the time. Baron's ( 1 99 1 )
model allowed growth whenever the skeleton was in ten-
sion due to internal pressure. This implies that during
periods of no growth, the pressure must be lower. But we
found that in well-fed, growing (Ebert, 1968) and starved,
possibly shrinking, urchins (Levitan, 1988; 1989). the
mean pressures were equal after the initial pressure surges
in the first 200-s traces (Table I).
The discrepancies between our observations and Bar-
COELOMIC PRESSURE IN URCHINS
433
on's ( 1991 ) assumptions have two possible implications:
that our specimens were abnormal, or that his assump-
tions do not reflect the pressure patterns in real urchins.
In the latter case, it may be that the spirit of the pneu
hypothesis is wrong, or that Baron's (1991) version does
not incorporate exactly the right assumptions. These pos-
sibilities can only be resolved by further experiments and
more refined theories.
At present, the most detailed predictions of urchin
shape, based on Baron's ( 199 1 ) tensile growth model, deal
only with regular urchins. A challenge to all models is the
great diversity of forms that must be generated, including
flattened sand dollars (Clypeasteroida), heart urchins
(Spatangoida), and the bizarre flask-shaped pourtalesiids
(Holectypoida).
Acknowledgments
This work was supported by University of California,
Davis, Agricultural Experiment Station Project no. 5134-
H and a U. C. Davis. Bodega Marine Laboratory Travel
Grant to O. Ellers: and a Natural Sciences and Engineering
Research Council of Canada Grant (#A4696) to M. Tel-
ford. We thank Bodega Marine Laboratory for use of their
facilities. We specially thank K. Brown for her organi-
zational help and H. Fastenau for diving to collect the
urchins and subsequently caring for and feeding them.
We thank J. Swanson and the staff of Keys Marine Lab-
oratory, Florida, for the use of their facilities and help in
collecting urchins. Thanks also to M. Martinez and K.
Driver who assisted in some of the experiments and to
D. Levitan who critically read the manuscript.
Literature Cited
Andrietti, F., M. D. Candia Carnevali, I. C. \\ilkie, G. Lanzavecchia,
G. Melone, and F. C. Celentano. 1990. Mechanical analysis of the
sea-urchin lantern: the overall system in Paracentrotus liviiitts. J
/.mil L< mil 220: 345-366.
Baron, C. J. 1988. Do mechanical forces explain patterns of growth
and form in sea urchins? A finite element analysis. Proc. Int. Echi-
noderms C'imf.. Victoria. Canada 1987: 786.
Baron, C. J. 1991. The structural mechanics and morphogenesis of
extant regular echinoids having rigid tests. Ph. D. Dissertation, Uni-
versity of California, Berkeley. 269 pp.
Candia Carnevali, M. D., F. Bonasoro, F. Andrictli, and G. Melone.
1990. Functional morphology of the peristomial membrane of reg-
ular sea urchins: structural organization and mechanical properties
in Paracentrotus li\idus. Pp. 207-216. in Echinoderm Research, De
Ridder. Duhois. Lahaye and Jangoux. eds. Balkema. Rotterdam.
Clark, R. B., and J. B. Cowey. 1958. Factors controlling the change
of shape of certain nemertean and turbellanan worms. J. Exp. Bio/.
35:731-748.
Dafni. J. 1983. Aboral depressions in the tests of the sea urchin Trip-
licates cf. gratilla (L.) in the gulf of Eilat. Red Sea. ./ Exp. Mar.
Bio/. Ecol. 67: 1-15.
Dafni. .1. 1985. Effect of mechanical stress on the calcification pattern
in regular echinoid skeletal plates. Proc. Int. Echinoderms Con]'.. Gal-
way 1984: 233-236.
Dafni, J. 1986. A biomechanical model for the morphogenesis of regular
echinoid tests. Paleabiology 12(2): 143-160.
Dafni, J., and J. Erez. 1982. Differential growth in Tripneustes gralilla
(Echinoidea). Proc. Int. Echinoderms Conf.. Tampa Bay 1981: 71-
75.
Ebert, T. A., 1968. Growth rates of the sea urchin Strongylocentrotus
purpuralus related to food availability and spine abrasion. Ecology
49: 1075-1091.
Ellers, O., and M. Telford. 1991. Forces generated by the jaws ofcly-
peasteroids (Echinodermata: Echinoidea). / Exp. Bio/ 155: 585-
603.
Fechter, H. 1965. Uber die Funktion der Madreporenplatte der Echi-
noidea. Z. I'tri-. Physiol. 51: 227-257.
Hanson, J. L., and G. Gust. 1986. Circulation of perivisceral fluid in
the sea urchin Lytechinus variegatus. Mar. BID/. 92(1): 125-134.
Hyman, L. H. 1955. The Invertebrates: Echinodermata. The Coelomale
Bilateria. Volume IV. McGraw-Hill, New York. 763 pp.
Jensen, M. 1985. Functional morphology of test, lantern, and tube
feet ampullae system in the flexible and rigid sea urchins (Echinoidea).
Proc. Int. Echinoderms Con/.. Galway 1984: 281-288.
Kier, \V. M., and A. M. Smith. 1990. The morphology and mechanics
of octopus suckers. Biol. Bull 178: 126-136.
Levitan, D. R. 1988. Density-dependent size regulation and negative
growth in the sea urchin Diadema aniillurum Philippi. Oecologia
76: 627-629.
Levitan, D. R. 1989. Density-dependent size regulation in Diadema
untillanun: effects on fecundity and survivorship. Ecology 70(5):
1414-1424.
Mortensen, Th. 1977. Handbook of the Echinoderms of the British
Isles Dr. W. Backhuys. Uitgever. Rotterdam. 47 I pp.
Moss, M. L., and M. M. Meehan. 1968. Growth of the echinoid test.
Acta Amitom. 64: 409-444.
Popov, E. P. 1976. Mechanics of Materials. 2nd ed. Prentice-Hall, En-
glewood Cliffs, NJ. 590 pp.
Royles, R., A. B. Sofoluwe, M. M. Baig, and A. J. Currie. 1980.
Behavior of underwater enclosures of optimum design. Strain Jan-
uary, 1980, pp. 12-20.
Seilacher, A. 1979. Constructional morphology of sand dollars. Pa-
leohiology 5(3): 191-221.
Smith, A. B. 1984. Echinoid Palaeobiology. George Allen and Unwin,
London. 190 pp.
Smith, A. M. 1991. Negative pressure generated by octopus suckers:
a study of the tensile strength of water in nature. / Exp. Biol. 157:
257-271.
Thompson, D. W. 1917. On Growth and Form. Cambridge University
Press, London.
Timoshenko, S., and S. VVoinowsky-Krieger. 1959. Theory of Plates
and Shells, 2nd ed. McGraw-Hill, New York. 580 pp.
Trueman, E. R. 1975. The Locomotion of Soft-bodied Animals. Edward
Arnold. London. 200 pp.
Vogel, S. 1981. Lite in Moving Fluids Willard Grant Press, Boston.
352 pp.
Vogel, S. 1988. Lite's Devices. Princeton University Press, Princeton,
NJ. 367 pp.
Voltzow, J. 1986. Changes in pedal intramuscular pressure corre-
sponding to behavior and locomotion in the marine gastropods fiif-
sycon conlranwn and Ha/iolis kamplschalkana Can. J Zool. 64( 10):
2288-2293.
\\ain«righl, S. A., \V. D. Biggs, J. D. Currey, and J. M. Gosline.
1976. Mechanical Design in Organisms Princeton University Press,
Princeton, NJ. 423 pp.
434
O. ELLERS AND M. TELFORD
Appendix 1
List of theoretical variables
Ap, pressure drop across a membrane
T, tension in the membrane
r, radius of curvature
T, . tangential tension in one direction in the membrane,
r,, radius of curvature associated with T,
T2, tangential stress in the direction perpendicular to T,
r>, radius of curvature associated with T:
pv, gravitational pressure, (not including atmospheric
pressure)
d, water depth
p, density of seawater
g, acceleration due to gravity
pd, dynamic pressure
u, speed of flow
f,,, vertical force exerted on the membrane by the lantern
weight and muscles
f,,,, is the tangential force in the membrane at the mem-
brane's attachment to the teeth
0, angle of membrane's attachment to the lantern (see Fig.
5b), same as tangential angle denned by fm
T. tension in the membrane
r,, radius of the central margin of the peristomial mem-
brane
r,,,,,, radius of curvature of the peristomial membrane
v, lantern protraction distance
h, horizontal distance from the central margin to the distal
margin of the peristomial membrane.
L, arc length of the peristomial membrane
4>, see diagram in Figure 5b.
0, v path — the combination of 0 and v used by the mem-
brane as it protracts
Reference: Bio/ Bull 182: 435-443. (June. 1992)
Hydrogen Sulfide Reduction of Symbiont Cytochrome
c552 in Gills of Solemya reidi (Mollusca)3
DAVID W. KRAUSbc, JEANNETTE E. DOELLERb, AND JONATHAN B. WITTENBERG
Department of Physiology and Biophysics, Albert Einstein College of Medicine.
1300 Morris Park Avenue. Bronx. New York. 10461
Abstract. The gill of the protobranch clam Solemya reidi
houses a dense population of intracellular symbiotic che-
moautotrophic sulfur-oxidizing bacteria that fix carbon
dioxide into sugars and supply the carbon nutrition of
the host. The gill is divided into a bacteriocyte (cells with
intracellular symbionts) domain and a domain of mito-
chondria-rich, symbiont-free ciliated cells. Optical spectra,
recorded separately from each domain, are dominated by
hemoglobin. Only oxygenated and deoxygenated hemo-
globin were detected in the gill. In sharp contrast to the
gill of the congener Solemya velum, ferric hemoglobin
sulfide was not detected, suggesting that this species, if
formed, is short lived. The spectral contribution of he-
moglobin may be cancelled or subtracted in difference
spectra. Difference spectra of each gill domain in nitrogen
minus the same tissue in air show a complement of re-
duced cytochromes. demonstrating that both symbiont
and mitochondrial cytochromes are reduced by endoge-
nous substrate. Difference spectra of the bacteriocyte do-
main exposed to hydrogen sulfide (air containing 1 .4 ton-
hydrogen sulfide minus air) show only the contribution
of reduced symbiont cytochrome t'55: . The extent of re-
duction increases monotonically with ambient /?H,s- sug-
gesting that, by analogy with some free-living sulfur-oxi-
dizing bacteria, cytochrome tV: is near the point of entry
of electrons into the symbiont electron transport chain.
Difference spectra of muscle or of the ciliated domain
under these same conditions show reduced cytochrome
csso< cytochrome /> and cytochrome oxidase, suggesting
Received 30 December 1991; accepted 2 March 1992.
* Part of this work was presented at the American Society of Zoologists
meeting. December. 1990.
b Present address: Department of Biology. University of Alabama at
Birmingham, Birmingham. Alabama. 35243.
c To whom correspondence should be sent.
that host mitochondria may accept electrons from hy-
drogen sulfide.
Introduction
Sulfide-oxidizing symbiotic associations between in-
vertebrate hosts and chemoautotrophic bacteria were first
recognized in members of the dense animal communities
found at deep ocean hydrothermal vents (reviewed in
Jones, 1985), and subsequently have been found in a
number of animal species inhabiting coastal and deep
sediments with disequilibrium mixtures of oxygen and
sulfide (reviewed in Southward, 1987; Fisher, 1990; Chil-
dress and Fisher, 1992). The bacterial symbionts oxidize
sulfide or other reduced compounds and fix carbon diox-
ide into nutrients that are translocated to the host (Fisher
and Childress, 1986). The gigantic size of some symbiont-
harboring animals attests to the effectiveness of the sym-
biotic association.
Symbionts in association with mollusks (clams, mus-
sels, and a snail) are housed intracellularly in specialized
cells — bacteriocytes — of the large and extensively modi-
fied gill. Within each cell they must be supplied with large
influxes of hydrogen sulfide and oxygen, but must be pro-
tected from excessive hydrogen sulfide that would inhibit
bacterial terminal oxidases and host mitochondrial cy-
tochrome oxidase (Nicholls, 1975; Wilson and Erecinska,
1978; see Somero et ai. 1989, and Childress and Fisher,
1992, for reviews). This condition must be met in the
naturally occurring steady state where hydrogen sulfide
entry is matched by utilization, and the cytoplasmic con-
centration of hydrogen sulfide. and possibly oxygen, is
probably very low (in the micromolar range; Childress,
1987; Wittenberg and Kraus, 1991). These cytoplasmic
concentrations of hydrogen sulfide and oxygen may not
be sufficient to support the fluxes of hydrogen sulfide and
435
436
D. W. KRAUS ET AL
oxygen to the symbiont, and we suggest that cytoplasmic
sulfide-reactive and oxygen-reactive hemoglobins may fa-
cilitate diffusion of their ligands through the cytoplasm
(Doeller et al, 1988; Kraus and Wittenberg, 1990; Wit-
tenberg and Kraus, 1991).
Hemoglobin is a nearly constant feature of symbioses
between mollusks and sulfide-oxidizing chemoautotraphic
bacteria (Wittenberg, 1985) and may reach concentrations
as high as 1.5 mM in the bacteriocyte (Kraus and Wit-
tenberg, 1990; for a possible exception see Dando et al..
1986). Two abundant, high oxygen affinity, cytoplasmic
hemoglobins have been isolated from the symbiont-con-
taining gills of Liicina peetinata (formerly named Pha-
coides pectinatus) (Read, 1962, 1965). One of these, he-
moglobin I. a monomer, reacts reversibly with oxygen
but also reacts rapidly and reversibly with hydrogen sulnde
in the presence of oxygen to form ferric hemoglobin sulnde
(Kraus and Wittenberg, 1990; Kraus et al.. 1990). It may-
be called "sulfide-reactive." The other, the "oxygen-re-
active" hemoglobin, is most probably an alpha:beta2 tet-
ramer made up of hemoglobins II and III (Kraus and
Wittenberg, 1990; Kemling et al.. 1991 ). It reacts revers-
ibly solely with oxygen. These two hemoglobins may de-
liver their respective ligands to the symbiont.
Two cytoplasmic hemoglobins occur in nearly equal
concentrations in the symbiont-harboring gill ofSolemya
velum, congeneric with Soleinya reidi of this study
( Doeller et al., 1983, 1988). Approximately half the total
hemoglobin within the bacteriocyte domain of living gill
filaments reacts reversibly to form ferric hemoglobin sul-
nde in the presence of sulfide and oxygen; the balance
remains oxygenated (Doeller et al, 1988). We infer that
the hemoglobins ofSolemya velum, like those of Liicinu.
may deliver their ligands hydrogen sulfide or oxygen to
the symbiont.
Here we present optical spectra of the symbiont-har-
boring bacteriocyte and the mitochondria-rich ciliated
domains of the gill of Solemya reidi, and compare these
spectra with those of additional symbiont-free tissues. We
find that, in contrast to the gill of the congeneric species
Solemya velum, and in contrast to purified Litcina pec-
nnata hemoglobin, the gill ofSolemya reidi never displays
the spectrum of ferric hemoglobin sulfide. Optical spectra
ascribed to the bacterial symbiont. weakly apparent in
other species, are seen with remarkable clarity in the gill
of Solemra reidi.
Materials and Methods
Animals
Solemya reidi individuals were collected by Van Veen
grab sampling at the Hyperion sludge outfall in Santa
Monica Bay, California, from a depth of 50-100 m. and
were maintained in cold (5-10°C) seawater. Animals were
3-5 cm long, and gill wet weight averaged 1.5 g. Experi-
ments were completed within 3 weeks of animal collec-
tion.
Optical spectrophotometry
Optical spectra were acquired with a Gary model 14
recording spectrophotometer equipped with a Gary scat-
tered transmission accessory and an Aviv digital data ac-
quisition and analysis system (Aviv Associates, Lakewood,
New Jersey). Optical spectra were recorded from 650 to
350 nm at 0.5 nm intervals. Experiments were performed
at room temperature (22-24°C). This is well above the
habitat temperature ofSolemya reidi from the collection
site(<10°C).
Optical spectra of gills and other tissue
Gills of Solemya reidi were excised and rinsed in 0.25
nm millipore-filtered seawater. Individual gill filaments
were cut from the central ligament of the gill. For some
experiments, filaments were divided along the chitinous
rod to separate the small ciliated mitochondria-rich do-
main from the larger bacteriocyte domain. Whole gill fil-
aments or filament domains, each about 40 nm thick (see
micrographs in Powell and Somero, 1985; Fisher and
Childress, 1986), were placed as a continuous overlapping
layer, two to three filaments or about 80 to 1 20 fim thick
on a gas-permeable membrane window (MEM 213, 25
nm thick. General Electric Corp., Schenectady. New
York). The layer was covered with a second membrane,
and the assembly placed in the previously described gas
perfusion cuvette (see Doeller et al.. 1988). Each gill prep-
aration was used for no longer than four hours, and for
each new experiment, filaments were freshly cut from the
gill. Other tissues, foot, adductor and pallial muscles, hy-
pobranchial gland, and nerve trunks or ganglia were ex-
amined as thin layer samples. Pallial muscles, if cut free,
contract and become thick. A useful preparation was ob-
tained by leaving a portion of the mantle with its fringing
pallial muscle attached to a fragment of the valve. A win-
dow cut in the valve allowed the light beam to pass through
the muscle. A single layer of parafilm (American Can Co..
Greenwich, Connecticut) was used to attenuate the Cary
reference beam and partially balance light scattering.
Gas delivery
Mixtures of air, oxygen, nitrogen, and carbon monoxide
were prepared using a Tylan mass flow controller (Carson,
California). Gas mixtures were humidified and passed
through the 5 ml spectrophotometer cuvette at a flow rate
of 100 ml/min. Hydrogen sulfide was added to the hu-
midified gas mixture from a glass syringe driven by a sy-
ringe pump (Harvard).
SYMBIONT CYTOCHROME f REDUCTION
437
Protocol
For each experiment, tissue hemoglobin was first ox-
ygenated and then deoxygenated by equilibration of the
tissue with air and nitrogen, respectively. Optical spectra
were recorded after each equilibration. The difference be-
tween these spectra was dominated by the contribution
of the difference: oxyhemoglobin minus deoxyhemoglo-
bin. The magnitude of this difference, together with an
estimate of the tissue thickness, provided an estimate of
hemoglobin concentration in each tissue sample. Subse-
quently, the samples of each tissue were exposed to mix-
tures of air and nitrogen in declining increments of 10%
air. from 100% air to 0% air, and spectra were recorded
at each step. The /'0, at which hemoglobin in each sample
was just detectably deoxygenated was noted and was used
in subsequent experiments with hydrogen sulfide to min-
imize spontaneous sulfide oxidation. This oxygen pres-
sure, which is influenced by hemoglobin oxygen affinity,
rate of oxygen consumption and tissue thickness, was
typically near 30 torr (20% air) for gill preparations and
near 90-105 torr (60-70% air) for pallial muscle prep-
arations.
Hydrogen sulfide concentration in the natural environ-
ment of Solemya reidi is large and variable, 0.1-3 mAI
hydrogen sulfide (Childress and Lowell, 1982). but the
animal may control the concentration of sulfide in the
ventilatory water current by changing the construction of
the burrow and its own placement within the burrow.
The partial pressure range of hydrogen sulfide used in
most these experiments (0.2-2 torr) is equivalent to 30-
300 n.M dissolved hydrogen sulfide (Millero. 1986). These
concentrations are near the range reported to support sul-
fide-dependent carbon fixation by Solemya velum gill (200
l*M: Cavanaugh. 1983a) or sulfide-dependent carbon
dioxide net uptake by Solemya reidi (50-200 nM; An-
derson et ai. 1987).
Hemoglobin isolati/ >n
Hemoglobin was purified following the general methods
of Schuder et ul. (1979) and Appleby et ai (1983). with
modifications introduced by Kraus and Wittenberg
(1990). The procedure involved extraction of the pow-
dered sample under an atmosphere of carbon monoxide
and argon, followed by molecular exclusion and ion ex-
change chromalography. All steps were carried out under
carbon monoxide-saturating conditions to maintain he-
moglobin in a carbon monoxide-ligated state until iso-
lation was complete. This minimizes oxidation of ferrous
Solemya hemoglobin and obviates cross-linking of he-
moglobin to itself and other tissue components.
Results
Hemoglobin identification and concentration
Cytoplasmic hemoglobin dominates the optical spectra
of the bacteriocyte and ciliated domains of the gill, as well
as of all other tissues examined, including foot, adductor
and pallial muscles, hypobranchial gland, and nervous
tissues. In each tissue, optical difference spectra of tissues
equilibrated with nitrogen minus the same tissue equili-
brated with air display well-resolved features at 412. 435,
540, and 580 nm, diagnostic for hemoglobin (the differ-
ence spectrum of the gill bacteriocyte domain is shown
in Fig. 1A). These maxima are the same as those in the
deoxyhemoglobin minus oxyhemoglobin difference spec-
trum of purified Solemya reidi hemoglobin. The concen-
tration of hemoglobin in several tissues (Table I) was cal-
culated from the difference in optical density at 435 nm
and 412 nm in these difference spectra, taking tissue
thickness estimated with a dissecting microscope and using
the extinction coefficient AEmM = 135, obtained from
the dominant fraction of hemoglobin isolated from So-
lemya reidi gills. Hemoglobin concentration calculated
from difference spectra reflects only the hemoglobin that
reversibly binds oxygen. Hemoglobin concentrations //;
situ are comparable with values obtained using hemoglo-
bin isolated from samples of whole gills and of remaining
tissues combined (Table I) and with the concentration of
hemoglobin reported by Powell and Arp (1989) for gills
ofSolenmi reidi ( 130 p.U).
Optical spectra of the hacteriocyte domain of the gill
Exposure to sulfide: hemoglobin. We now examine the
spectral changes induced by exposing the bacteriocyte do-
main of the gill to hydrogen sulfide. The bacteriocyte do-
main was first equilibrated with 20%- air, sufficient oxygen
to just saturate the hemoglobin in these samples, then
hydrogen sulfide (about 0.2 torr) was added to the hu-
midified gas. During equilibration with sulfide. absorbance
was recorded continuously at either 420 nm or 430 nm,
wavelengths near the maxima of ferric hemoglobin sulfide
or deoxyhemoglobin. respectively. Absorbance exhibited
a monotonic change with time, reaching an asymptote
within 80 ± 30 s (n = 6). At steady state, an optical spec-
trum was recorded. Sulfide partial pressure was then in-
creased incrementally to 2 torr, with the same procedure
repeated at each step. Difference spectra of gills in 20%.
air containing sulfide minus gills in 20% air alone did not
produce any feature that could be ascribed to sulfide li-
gation to hemoglobin (see Doeller et ai. 1988: Kraus and
Wittenberg. 1990). but instead revealed an apparently
single spectral entity with characteristics of a reduced mi-
nus oxidized cytochrome t'55:. discussed below. Thus, we
cannot detect the formation of ferric hemoglobin sulfide
in the bacteriocyte domain of the gill of Solemya reidi.
438
D. W. KRAUS ET AL.
435
0.025 OD
412
424
552
0.005 OD
424
0.05 OD
D
422
550
0.02 OD
0.01 3 OD
0.008 OD
419
550
0.01 OD
0.005 OD
407
400 450 500 550 600 4OO
Wavelength, run
450
500
550
600
Figure 1. Optical difference spectra of living Solcmya rciiti tissues. Designated wavelength maxima are
discussed in detail in text. Traces in the visible region have been amplified four-fold in Figures 1A-C. 2.5-
fold in Figure ID. and two-fold in Figure IE.
A. Bacteriocyte domain of gill filaments equilibrated with nitrogen minus the same sample equilibrated
with air. The contribution of deoxyhemoglobin minus oxyhemoglobin dominates the difference spectrum,
with diagnostic maxima at 412, 435. 540. and 580 nm.
B. Bacteriocyte domain of gill filaments equilibrated with 20'? air containing 1.4 torr sulfide minus the
same sample equilibrated with 20% air. The dominant spectral species is identified as a reduced minus
oxidized cytochrome r552, ascribed to bacterial cytochrome c'552.
C. Bacteriocyte domain of gill filaments equilibrated with nitrogen minus the same sample equilibrated
w ith 20'7 air, from which a difference spectrum of purified deoxyhemoglobin minus oxyhemoglobin, computed
to be equivalent to the hemoglobin content in gills, was subtracted. The wavelength maxima indicated were
confirmed in the second derivative of this difference spectrum. Features are ascribed to bacterial cytochrome
(55: and to a cytochrome b (566 nm). The feature at 586 nm is not identified.
SVMBIONT CYTOCHROME c REDUCTION
439
Table I
.•i/>rir<i\inhilc hemoglobin concentration in tissues of Solemya reidi.
determined in living tissues and from hemoglobin extruded from tissue
I issue
Hemoglobin
concentration, ,uM
Living tissue3
bacteriocyte domain of gills
ciliated domain of gills
pallial muscle
hypobranchial gland
Hemoglobin extraction
whole gill
combined symbiont-free tissue (foot.
pallial muscle, hypobranchial
gland )
450 ± 90(21)"
120 ± 15 (4)
330(2)
540 (2)
180
170
* Tissue thickness was estimated with a dissecting microscope at 80
iim for bacteriocyte domain. 125 ^m for ciliated domain. 300 ^m for
pallial muscle and 100 ^m for hypobranchial gland.
b Numbers are given as average ± standard deviation (number of rep-
etitions).
Exposure to sulfide: cytochrome c*,f:. The optical con-
tribution of hemoglobin did not change in the presence
of hydrogen sulh'de, either aerobically or anaerobically.
Consequently, the spectral contribution of oxyhemoglobin
cancels in the difference spectrum: gills in 20% air con-
taining sulnde minus gills in 20% air alone. The remaining
spectrum in the bacteriocyte domain (Fig. IB), charac-
terized by sharp features at 405, 424, 520, and 552 nm,
is identified as the difference: reduced minus oxidized cy-
tochrome 1-552 (Pettigrew and Moore, 1987). This is un-
equivocally distinguishable from mitochondria! cyto-
chrome c with maxima in the direct reduced spectrum at
520 and 550 nm (see Figs. ID, E; Pettigrew and Moore,
1987). and may be ascribed to the symbiont (see Discus-
sion). Thus, hydrogen sulnde causes reduction of symbiont
cytochrome 1-552 in the gill of Solemya rculi, without de-
tectable change in other heme proteins. We note that re-
duction of cytochrome c^2 was not detected in gills ex-
posed to 300 n.M thiosulfate in seawater (data not shown).
This may reflect the lack of uptake of thiosulfate into
bacteriocyte cytoplasm.
I
•o
i.oo-
0.75-
0.50-
0.25-
0.00
(3)
(I)
1 1 r—
0123
pH2S, torr
Figure 2. Relative reduction of cytochrome (.551 in the bacteriocyte
domain of living Solemya reidi gill filaments as a function of ambient
partial pressure of hydrogen sulnde in air. Relative reduction is calculated
from difference spectra as described in text. Half-reduction of cytochrome
f551 occurs near 1.4 torr/>H,s- Numbers in parentheses represent number
of experiments.
The relative extent of reduction of cytochrome 1-552 in
the bacteriocyte domain of aerobic gills increases mono-
tonically with ;>H,s from zero m the absence of sulfide to
a limit near 7 torr pHlS (Fig. 2). Relative reduction was
calculated from the magnitude of the optical density dif-
ferences 424 nm minus 405 nm and, separately, 552 nm
minus 540 nm in difference spectra similar to Figure IB
(gills in 20% air containing sulnde minus gills in 20% air),
normalized to a constant amount of hemoglobin in the
optical path so as to allow comparison of samples of dif-
ferent thickness. Hemoglobin content was estimated from
difference spectra similar to Figure 1A (gills in nitrogen
minus gills in air) obtained from the same tissue sample.
Reduction of cytochrome c\52 wa$ taken as maximal at
/>H,S = 1 torr and minimal in the absence of hydrogen
D. Ciliated domain of gill filaments equilibrated with nitrogen minus the same sample equilibrated with
40% air, from which a difference spectrum of purified deo.xyhemoglobin minus oxyhemoglobin. computed
to be equivalent to the hemoglobin content in gills, was subtracted. The wavelength maxima indicated were
confirmed in the second derivative of this difference spectrum. Features are ascribed to mitochondria! cy-
tochrome (-550 and cytochrome />
E. Ciliated domain of gill filaments equilibrated with 40% air containing 1.4 torr hydrogen sulfide minus
the same sample equilibrated with 40% air. The wavelength maxima indicated were confirmed in the second
derivative of this difference spectrum. Features are ascribed to mitochondria! cytochrome r550, cytochrome
/). and cytochrome oxidase.
F. Pallial muscle equilibrated with 70% air containing 1.4 torr hydrogen sulfide minus the same sample
equilibrated with 70%> air. Features are ascribed to mitochondria! cytochromes.
440
D. W. KRAUS ET AL
sulfide. Half reduction of cytochrome (.'552 occurred near
1 .4 torr />H,S-
Exposure to nitrogen: cytochromes. The spectral con-
tribution of reduced cytochromes in the optical spectrum
of the gill is obscured by the spectral contribution of he-
moglobin. Accordingly, the hemoglobin content of each
sample was estimated and the equivalent difference spec-
trum: deoxyhemoglobin minus oxyhemoglobin, was sub-
tracted from the difference spectrum: gills in nitrogen mi-
nus gills in air. The remaining spectral contribution (Fig.
1C) is dominated by bacterial cytochromes because mi-
tochondria are sparse. The features at 552 and 424 nm
correspond to the bacterial cytochrome c^2- The shoulder
near 566 nm may tentatively be attributed to a cyto-
chrome h. The feature at 586 nm remains unidentified.
Exposure to nitrogen plus sulfide: cytochrome c'_w.
Maximal reduction of cytochrome c^2 was observed in
gills exposed to nitrogen containing 1 .4 torr hydrogen sul-
fide, demonstrated in the difference spectrum between
this condition and gills in air (data not shown). Reduction
of cytochrome c^ in gills exposed to nitrogen alone (Fig.
1C) represented only 81 ± 6% (n = 10) of maximum. It
follows that symbiont cytochrome t'55: is largely reduced
in nitrogen alone, but is reduced still further in the pres-
ence of hydrogen sulfide. We note that the optical density
differences (424 nm minus 405 nm and 552 nm minus
540 nm) are the same in gills exposed to 1 .4 torr hydrogen
sulfide in the presence of nitrogen and in gills exposed to
7 torr hydrogen sulfide in air, indicating that maximal
reduction was achieved in each instance.
Optical spectra of the ciliated domain of the gill
Exposure to nitrogen: cytochromes. Cytoplasmic he-
moglobin, at a concentration roughly one third that of
the bacteriocyte domain (Table I), dominates the optical
spectrum of the symbiont-free ciliated domain of the gill.
A reduced minus oxidized difference spectrum (Fig. ID)
was constructed by subtracting the expected spectral con-
tribution of cytoplasmic hemoglobin from the difference
spectrum: gills in nitrogen minus the same tissue in air
or 40% air. Clearly resolved spectral features at 422, 520,
and 550 nm may be ascribed to mitochondrial cytochrome
c, identified by comparison with the difference of reduced
minus oxidized horse heart cytochrome c, which has fea-
tures at 419, 520. and 550 nm (Pettigrew and Moore.
1987). The 550 nm feature in the ciliated domain differ-
ence spectrum is consistently distinct from the 552 nm
feature seen in the bacteriocyte domain difference spec-
trum (Fig. 1 B). The shoulder at 565 nm may be reasonably
ascribed to mitochondrial cytochrome b (ubiquinone-cy-
tochrome c1 oxidoreductase).
Exposure to sulfide: cytochromes. We next examine the
spectral change produced by exposing the air-equilibrated
ciliated domain to 1.4 torr hydrogen sulfide. The differ-
ence spectrum (Fig. IE) is very similar to that of partially
reduced minus oxidized mitochondria. Features at 419,
520, and 550 nm may again be ascribed to cytochrome
c. Features at 446 and 602 nm may be ascribed to cyto-
chrome oxidase. Small features near 430 and 565 nm may
be ascribed to cytochrome /'. All of these features become
more prominent with increased hydrogen sulfide to 2.8
torr (data not shown).
Exposure to nitrogen plus sulfide: cytochromes. Further
reduction of mitochondrial cytochromes by hydrogen
sulfide in nitrogen compared to nitrogen alone was not
observed. The optical difference: gills in nitrogen con-
taining 1.4 torr hydrogen sulfide minus gills in nitrogen
alone was relatively featureless and exhibited no peaks
ascribed to mitochondrial cytochromes (data not shown).
Optical spectra of pallial muscle
Exposure to sulfide: cytochromes. The pink-colored
pallial muscle is another example of symbiont-free tissue
with cytoplasmic hemoglobin. In the difference spectrum
of a thin piece of pallial muscle equilibrated with 70% air
containing 1 .4 torr hydrogen sulfide minus the same sam-
ple in 70% air alone, the spectrum of hemoglobin cancels
and the remaining difference spectrum suggests reduction
of a full complement of mitochondrial cytochromes: cy-
tochrome r55o, cytochrome b and cytochrome oxidase
(Fig. IF).
Discussion
The protobranch mollusk Solemya rcidi lives in bur-
rows in strongly reducing sediments, with hydrogen sulfide
concentrations reaching 3 mM. It is quite mobile and
may seek appropriate sulfide concentrations (Reid, 1980:
Childress and Lowell. 1982). There is no sulfide-binding
protein in the circulating blood and the sulfide concen-
tration in gills of freshly captured specimens is close to
environmental (Childress. 1987). suggesting that sym-
bionts take up hydrogen sulfide and oxygen directly across
the gill. Without question, symbionts use hydrogen sulfide.
Sulfide stimulates oxygen and carbon dioxide consump-
tion of Solemya rcidi (Anderson el a/., 1987) and carbon
fixation in isolated gills of So/emya velum (Cavanaugh.
1983a). We note that on a per sulfur basis, sulfide is much
more effective in stimulating carbon dioxide fixation than
thiosulfate: approximately 14-fold more effective in whole
Solemya reidi (recalculated from Anderson et a/.. 1987),
and 6-fold more effective in isolated Solemya velum gills
(recalculated from Cavanaugh, 1983a).
The gill of Solemya reidi is comprised of a few hundred
individual filaments held together at a central ligament
and arranged in parallel somewhat like pages of a book.
Each filament is divided into two major domains by a
chitinous skeletal rod located near the outer edge: a rel-
SYMBIONT CYTOCHROME C REDUCTION
441
atively small outer domain of mitochondria-rich ciliated
cells and a much larger inner domain comprised largely
of bacteriocytes (Yonge. 1939: Reid. 1980: Powell and
Somero. 1985). Ciliated cells near the outer edge drive
the flow of water between filaments and across the face
of the inner domain. Electron micrographs of Solemya
rcidi (Powell and Somero. 1985) and of the related species
Solcmya velum (Cavanaugh, 1983b; Doeller. 1986) show
that cells in the ciliated domain are densely packed with
mitochondria and are free of bacteria. Conversely, bac-
teriocytes are densely packed with bacteria; mitochondria
are sparse. Separation of the two cell types occurs also in
gills of several lamellibranch mollusks of the family Lu-
cinidae (Giere, 1985). including Lucina pectinata (J. B.
Wittenberg, unpub.). In this study we take advantage of
the separation of bacteriocyte and mitochondria-rich do-
mains to study each separately.
Cytochromes of the bacterial partner are singularly well
resolved in optical spectra of the bacteriocyte domain of
the Soicnivu rcidi gill. Difference spectra of the bacterio-
cyte domain exposed to a low concentration of hydrogen
sulfide in the presence of air are dominated by the spectral
contribution of a hemeprotein characterized by a prom-
inent alpha-band centered at 552 nm and identified as a
t-t\pe cytochrome by the positions of its wavelength
maxima, 424. 520, and 552 nm (Fig. 1 B). It may be called
cytochrome f55:. Cytochromes c characterized by maxima
between 551 and 553 nm are conspicuous components
of many sulfur-oxidizing bacteria (Pettigrew and Moore,
1987). They are not known in eukaryote tissues and are
easily distinguished from mitochondria! cytochrome c-,^,
observed, for instance, in the ciliated domain of the gill
and the pallial muscle (Fig. ID, E. and F). Narrow spectral
bandwidths in the observed spectrum (Fig. IB) and the
absence of features not related to cytochrome cS52 (other
than a small perturbation near 600 nm) suggest strongly
that hydrogen sulfide. in the presence of oxygen, has re-
duced a single spectrally demonstrable species, bacterial
cytochrome r55:. The extent of this reduction is a mono-
tonic function of ambient />H,s (Fig. 2). suggesting, but by
no means proving, that hydrogen sulfide is the immediate
or near immediate reductant for symbiont cytochrome
c'5j:. Reduction of Cytochromes by sulfide might proceed
by direct electron transfer (Wilson and Erecinska. 1978)
as it does in the free-living bacterium Thiobacillus deni-
trificans (Sawhney and Nicholas, 1978), or it could be
mediated by flavocytochromes c (reviewed in Pettigrew
and Moore. 1987). Reduction ofSolemya rcidi symbiont
cytochrome 1-552 finds a strong parallel in the free-living
sulfur-oxidizing bacteria, where soluble c-type cyto-
chromes occurring in the periplasmic space are considered
the main point of transfer of electrons from external re-
duced sulfur compounds into the bacterial cytochrome
chain (Kelly. 1982. 1985. 1988; Pettigrew and Moore,
1987).
Cytochrome r55: in the bacteriocyte domain of the gill
is also reduced under anaerobic conditions in the absence
of hydrogen sulfide. In this case, a cytochrome b and per-
haps other Cytochromes are reduced as well (Fig. 1C), and
we cannot define an unique path of electron flow. Nor do
we assert that the cytochrome r552 reduced anaerobically
in the absence of sulfide is the same species as that reduced
by sulfide. Our data indicate that symbiont cytochrome
c552 is largely reduced by endogenous substrate under an-
aerobic conditions, but exhibits further reduction by hy-
drogen sulfide under these conditions. We find a possible
analogy in the thiosulfate-oxidizing bacterium Thioba-
cillus versutus (A2). which has two r-type Cytochromes.
£•551 and t's52.5. each with two separately titratable oxida-
tion/reduction centers with widely different midpoint po-
tentials (Lu and Kelly. 1984; Lu et ai, 1984). Transfer of
electrons at two different potentials from thiosulfate to
cytochrome c- of the bacterial electron transport chain is
mediated by this multi-heme complex (Lu and Kelly,
1984; Lu et ai. 1984). The additional reduction of sym-
biont cytochrome c^2 in Solemya rcidi gills by hydrogen
sulfide. over and above that in nitrogen alone, indicates
that cytochrome 1-552 may have two oxidation/reduction
centers as well.
The symbiont-free ciliated domain of the gill, made
anaerobic, displays the expected spectrum of the reduced
mitochondria! electron transport chain, with features as-
cribable to cytochrome c. cytochrome b, and cytochrome
oxidase (Fig. ID). The ultimate reductant in this instance
must be endogenous substrate.
The mitochondria-rich tissues, pallial muscle with sin-
gularly well-resolved spectra and the ciliated domain of
the gill, exposed to hydrogen sulfide in the presence of
air, exhibit spectra of reduced mitochondria, once again
with features ascribed to cytochrome c. cytochrome />.
and cytochrome oxidase (Fig. IE. F). A simple explana-
tion is that hydrogen sulfide is serving as the reductant
for the respiratory chain. Indeed, isolated gill mitochon-
dria from Solemya rcidi are known to oxidize sulfide with
production of ATP (Powell and Somero. 1986; O'Brien
and Vetter. 1990). The inference from these studies is that
electrons from hydrogen sulfide enter the mitochondria!
electron transport chain at the level of cytochrome c, with
oxidative phosphors lation only at the cytochrome oxidase
site (Complex IV). An alternative explanation of our ob-
servations is that hydrogen sulfide may bind to and inhibit
cytochrome oxidase. with a spectral signature not easily
distinguished from normal oxidation/reduction (Wilson
and Erecinska. 1978). This also would lead to net observed
cytochrome reduction, in this instance by endogenous
substrate. Possibly both processes occur simultaneously.
442
D. W. KRAUS ET AL
Hemoglobin, presumably located in the host cell cy-
toplasm, is abundant in the symbiont-containing bacte-
riocytes of the Solemya reidi gill and occurs as well in the
symbiont-free ciliated domain of the gill, foot, pallial and
adductor muscles, hypobranchial gland, and nervous tis-
sue (Table I). Hemoglobin concentration in these tissues
is comparable to the myoglobin content of many hard-
working muscles (Schuder ct a!.. 1979). The concentration
of hemoglobin in the bacteriocyte domain of the gill is in
the upper part of the range reported for other symbiont-
housing molluscan gills (Wittenberg, 1985).
Only oxygenated and deoxygenated hemoglobin were
detected in the bacteriocyte domain of the Solemya reidi
gill and in other tissues (e.g.. Fig. 1 A). Ferric hemoglobin
sultide, that is ferric hemoglobin with sulfide ligated to
the heme iron atom in the distal position, was not detected
under any conditions. This stands in sharp contrast to
the behavior of hemoglobin in the living gill of the con-
generic species Solemya velum, where about half of the
gill hemoglobin is rapidly and reversibly converted to ferric
hemoglobin sulfide when the gill is exposed to low con-
centrations of hydrogen sulfide in aerated seawater
(Doeller et al.. 1988). The behavior of hemoglobin in the
Solemya reidi bacteriocyte also stands in contrast to the
reaction of the "sulfide-reactive" hemoglobin, Hb I, iso-
lated from the symbiont-containing gill of the lucinid clam
Lueina peetinata (Kraus and Wittenberg. 1990). Oxygen-
ated Lueina Hb I reacts rapidly with hydrogen sulfide at
micromolar concentration to form ferric hemoglobin sul-
fide. We propose that in these three symbioses. sulfide-
reactive gill hemoglobin functions to deliver either hy-
drogen sulfide or reducing equivalents to the symbiont.
As one working hypothesis, we offer that the observed
large difference in the steady-state concentration of ferric
hemoglobin sulfide in the two Solemya gills reflects very
different rates of chemical reaction. Extraordinary slow
dissociation of sulfide from ferric Lueina hemoglobin sul-
fide (koff= 2 X 10~4s~'; implying a turnover time of 5000
s) suggests that delivery of sulfide cannot be achieved by
simple dissociation of the ligand. Instead, we suggest re-
duction of ferric hemoglobin sulfide near the peribacterial
membrane surface may precede ligand delivery (Kraus
and Wittenberg. 1990; Wittenberg and Kraus. 1991 ). We
consider that this latter step may be rate-limiting in the
gill of Solemya velum but not of Solemya reidi. Reduction
of the major hemoglobin of the Solemya reidi gill, when
ferric, is very much more rapid than the corresponding
reduction of ferric Lueina hemoglobin sulfide in the pres-
ence of excess sulfide (Kraus and Doeller, unpub.). Per-
haps rapid reductive removal of ferric hemoglobin in the
Solemya reidi gill prevents accumulation of ferric he-
moglobin sulfide in the tissue. As an alternative hypoth-
esis, we offer that Solemya reidi gill hemoglobin may
function in the transfer of reducing equivalents from sul-
fide to symbiont. This function is suggested by the rapid
reduction of ferric Solemya reidi hemoglobin by hydrogen
sulfide and by the lack of formation of ferric Solemya
reidi hemoglobin sulfide in vitro (under the same condi-
tions as those that lead to the formation of ferric Lueina
hemoglobin sulfide; Kraus and Doeller. unpub.).
In summary, bacterial symbionts of the Solemya gill
use hydrogen sulfide as the sole environmental source of
reducing equivalents. We show here that symbiont bac-
terial cytochrome cx52 is extensively reduced when the gill
is exposed to hydrogen sulfide, and that hydrogen sulfide
is the immediate or near immediate reductant for this
cytochrome. This finds a strong parallel in free-living, sul-
fide-oxidizing bacteria that are considered to accept elec-
trons from reduced sulfur compounds at the level of cy-
tochrome c supporting oxidative phosphorylation at the
level of the terminal oxidase and reverse electron flow to
NAD (reviewed in Kelly, 1982, 1985, 1988; Pettigrew and
Moore, 1987). Only ox\ hemoglobin and deoxyhemoglo-
bin are detected in the gill of Solemya reidi. in sharp con-
trast to the congener So/emya velum where ferric hemo-
globin sulfide constitutes about half of the hemoglobin in
gills exposed to hydrogen sulfide and oxygen. Solemya
reidi gill hemoglobin may participate in the symbiosis by
rapid formation and reduction of ferric hemoglobin sul-
fide, or by transfer of electrons from sulfide to symbiont.
Acknowledgments
This work was supported in part by research grants
DMB 87-03328 and DCB 90-17722 from the National
Science Foundation (to JBW). in part by USPHS research
grant HL 19299 (to Dr. B. A. Wittenberg), and in part by
the University of Alabama at Birmingham Faculty Re-
search Grant (to DWK). Animals used in this study were
collected on a cruise of the Point Sur. supported by re-
search grant OCE 86-10514 from the National Science
Foundation (to Dr. J. J. Childress). J. B. Wittenberg is a
Research Career Program Awardee 1-K6-733 of the
United States Public Health Service, National Heart, Lung
and Blood Institute. We extend special thanks to Drs.
C. R. Fisher and J. J. Childress for the invitation to DWK
and JED to participate in the Solemya reidi collecting
cruise, to the captain and crew of the Point Sur for their
help in the collection of Solemya reidi. to Dr. R. G. B.
Reid for supplying specimens of Solemya reidi for pre-
liminary experiments, and to Dr. B. A. Wittenberg for
continuing discussions.
Literature Cited
Anderson, A. E., J. J. Childress, and J. A. Favuzzi. 1987. Net uptake
of CO2 driven by sulphide and thiosulfate oxidation in the bacterial
symbiont-containing clam S/I/CIHYU reidi. J Exp. Biol. 133: 1-31.
SYMBIONT CYTOCHROME < REDUCTION
443
Vppleb). C. A., J. D. Ijepkema. and M. J. I rinick. 1983. Hemoglobin
in a nonleguminous plant. Paras/ioma: possible genetic origin and
function in nitrogen fixation. Science 220: 951-953.
Ca>anaugh, ('. M. 19833. Symbiotic chemoautotrophic baeteria in
marine invertebrates from sulphide-rich habitats, \atiire 302: 58-
61.
i i>.iiuiii;li. C. M. I983b. Chemoautotraphic bacteria in marine in-
vertebrates from sulnde-rich habitats: a new symbiosis. Pp. 699-708
in Endocylobiology, I '<>/. //.. Intracelhi/ar Space as Oligogenetic Eco-
system, H. E. A. Schenk and W. Schvvemmler. eds. Walter de Gruyler
and Co.. Berlin.
Childress, J. J. 1987. Uptake and transport of sulfide in marine in-
vertebrates. In Comparative Physiology: Life in H'ulcr ami on Land.
P. Dejours. L. Bolis. C. R. Taylor, and E. R. Weibel, eds. Fidia
Research Series 9: 231-239. Liviana Press. Padova.
Childress, J. J., and C. R. Fisher. 1992. The biology of hydrothermal
vent animals: physiology, biochemistry and autotrophic symbioses.
Oceanogr. Mar BID/. Ann. Rev 30: in press.
Childress, J. J., and \V. Lowell. 1982. The abundance of a sulfide-
oxidizing symbiosis (the clam Solemya reidi) in relation to interstitial
water chemistry. Eos 63: 957.
Dando, P. R.. A. J. Southward, and E. C. Southward.
1986. Chemoautotrophic symbionts in the gills of the bivalve mol-
lusc Lucinoma borealis and the sediment chemistry of its habitat.
Proc. R Soe. Loud. B 227: 227-247.
Dueller. J. E. 1986. A study of the gill hemoglobin in the nearly gutless
bivalve Solemya velum Say (Protobranchia). Ph.D. Dissertation,
Clemson University, SC.
Doeller, J. K... D. W. Kraus, and J. M. Colaeino. 1983. The presence
of hemoglobin in Solemya IV/H/H (Bivalvia. Protobranchia). Am. Zoo/.
23: 976.
Doeller, J. E., D. \V. Kraus, J. M. Colaeino, and J. B. Wittenberg.
1988. Gill hemoglobin may deliver sulnde to bacterial symbionts
of Solemya velum (Bivalvia, Mollusca). Bio/. Bull. 175: 388-396.
Fisher, C. R. 1990. Chemoautotrophic and methanotrophic symbioses
in marine invertebrates. Rc\: Aquatic Sei. 2: 399-436.
Fisher, C. R., and J. J. Childress. 1986. Translocation of fixed carbon
from symbiotic bacteria to host tissues in the gutless bivalve Solemya
reidi. Mar. Biol. 93: 59-68.
Giere, O. 1985. Structure and position of bacterial endosymbionts in
the gill filaments of Lucinidae from Bermuda (Mollusca. Bivalvia).
ZooniorpholoKy 105: 296-301.
Jones, M. I,, (ed). 1985. Hydrothermal vents of the eastern Pacific: an
overview. Bull. Biol. Soc. Hash. 6: 1-547.
Kelly, D. P. 1982. Biochemistry of the chemolithotrophic oxidation
of inorganic sulphur. Phil. Trans. R Soe. Loud. B 298: 499-528.
Kelly, D. P. 1985. Physiology of the thiobacilli: elucidating the sulphur
oxidation pathway. Microbiol. Sei. 2: 105-109.
Kelly, D. P. 1988. Oxidation of sulfur compounds. In The .\nnyen
and Sulfur Cycles. Cole, J. A. and S. J. Ferguson, eds. Symp. Soc
Gen. Mierohiol. 42: 65-98.
K, ml. ii-. V, D. W. Kraus, J. D. Hockenhull-Johnson, J. B. Wittenberg,
S. N. Vinogradov, and P. Martin. 1991. Crystalli/ation of a complex
of hemoglobin components II and III of the symbiont-harboring clam
Lucina peetinala ./. .\fol. Biol. 222: 463-464.
Kraus, D. W., and J. B. Wittenberg. 1990. Hemoglobins of the Lucina
peclinata/bactena symbioses. I. Molecular properties, kinetics and
equilibna of reactions with ligands. ./. Biol. Chem 265:16,043-16,053.
Kraus. D. W., J. B. Wittenberg, L. Jing-Fen, and J. Peisach.
199(1. Hemoglobins of the Lucina pectinata/bacteria symbioses. II.
An electron paramagnetic resonance and optical spectral study of
the feme proteins. J Biol Chem. 265: 16.054-16.059.
Lu, W.-P., and D. P. Kelly. 198-1. Purification and characterization of
two essential cytochromes of the thiosulphate-oxidizing multi-enzyme
system from Thiobacillus A2 ( Thiobacillus versutus). Biochun. Bio-
phys. Ada 765: 106-1 17.
Lu, W.-P., R. K. Poolc. and D. P. Kelly. 1984. Oxidation-reduction
potentials and spectral properties of some cytochromes from Thio-
bacillus versulus (A2). Biochim. Biophys. Ada 767: 326-334.
Millero, F..1. 1986. The thermodynamics and kinetics of the hydrogen
sulfide system in natural waters. Mar. Chem. 18: 121-147.
Nicholls, P. 1975. The effect of sulphide on cytochrome aa3 isosteric
and allosteric shifts of the reduced alpha-peak. Biochim. Biophys
Ada 396: 24-35.
O'Brien, J., and R. D. Vetter. 1990. Production of thiosulphate during
sulphide oxidation by mitochondria of the symbiont-containing bi-
valve Solemya reidi. J. E.\p. Biol. 149: 133-148.
Pettigrew, G. W., and G. R. Moore. 1987. Cytochromes c. Biological
Aspects. Springer- Verlag, NY, Pp. 1-282.
Powell, M. A., and A. J. Arp. 1989. Hydrogen sulfide oxidation by
abundant nonhemoglobin heme compounds in marine invertebrates
from sulnde-rich habitats. J. E.\p. Zoo/ 249: 121-132.
Powell, M. A., and G. N. Somero. 1985. Sulfide oxidation occurs in
the animal tissue of the gutless clam. Solemya reidi. Biol Bull 169:
164-181.
Powell, M. A., and G. N. Somero. 1986. Hydrogen sulfide oxidation
is coupled to oxidative phosphor} lation in mitochondria of Solemya
reidi Science 233: 563-566.
Read, K. R. H. 1962. The hemoglobin of the bivalved mollusc. Pha-
coides peclinatus (Gmelin). Biol Bull. 123:605-617.
Read, K. R. H. 1965. The characterization of the hemoglobins of the
bivalve mollusc Phacoides peclinatus (Gmelin). Coinp. Biochem.
Physiol. 15: 137-158.
Reid, R. G. B. 1980. Aspects of the biology of a gutless species of So-
lemya (Bivalvia: Protobranchia). Can. J. Zool. 58: 386-393.
Sawhney, \ ., and D. J. D. Nicholas. 1978. Sulnde-lmked nitrite re-
ductase from Thiobacillus denitrificans with cytochrome oxidase ac-
tivity: purification and properties. / Gen. Mierohiol. 106: 1 19-128.
Schuder, S., J. B. Wittenberg, B. Haseltine. and B. A. Wittenberg.
1979. Spectrophotometric determination of myoglobin in cardiac
and skeletal muscle: separation from hemoglobin by subunit-exchange
chromatography. Anal. Biochem. 92: 473-481.
Somero, G. N., J. J. Childress, and A. E. Anderson. 1989. Transport,
metabolism, and detoxification of hydrogen sulfide in animals from
sulnde-rich marine environments. Rev. Aquatic Sei. I: 591-614.
Southward, E. C. 1987. Contribution of symbiotic chemoautotrophs
to the nutrition of benthic invertebrates. Pp. 84-1 16 in Microbes in
the Sea. M. A. Sleigh, ed. Ellis Horwood. Chichester.
Wilson, D. F., and M. Erecinska. 1978. Ligands of cytochrome c ox-
idase. Methods En:ymol. 53: 191-201.
Wittenberg, J. B. 1985. Oxygen supply to intracellular bacterial sym-
bionts. In Hvdrolhermal I 'cuts o/ the Eastern Pacific: An Overview,
M. L. Jones, ed. Bull. Biol Soc II Wi 6: 1-547.
Wittenberg. J. B.. and D. W. Kraus. 1991. Hemoglobins of eukaryote/
prokaryote symbioses. Pp. 323-330 in Structure and Function of In-
vertebrate Oxvgen Carriers. S. N. Vinogradov and O. H. Kapp. eds.
Springer- Verlag, New York.
Yonge, C. M. 1939. The protobranchiate molluscs: a functional inter-
pretation of their structure and evolution. Phil Trans. Roy Soc.
Lond. B 230: 79- 147.
Reference: Biol Bull 182: 444-453. (June, 1992)
Oxygen- and Nitrogen-Dependent Sulfur Metabolism
in the Thiotrophic Clam Solemya reidi
DAVID B. WILMOT1 AND RUSSELL D. VETTER
Marine Biology Research Division, 0202, Scripps Institution of Oceanography.
University of California, San Diego, La Jolla, California 92093
Abstract. We investigated aerobic and anaerobic thio-
trophic metabolism by the gutless clam Solemya reidi and
its intracellular symbiotic bacteria. Mean environmental
sulfide concentrations in porewater next to animals varied
from a high of 888 /uA/ to a low of 17 ftM in different
sediment samples, while mean thiosulfate concentrations
were very low (1-13 ^Af). The blood of freshly collected
clams contained up to 300 /uA/ thiosulfate but little sulnde
(<12 nAI). In experimental incubations, clams were able
to take up thiosulfate, yet under no conditions could the
animals concentrate thiosulfate above external concen-
trations. Thiosulfate accumulation in the blood during
incubations was the result of aerobic but not anaerobic
sulfide oxidation by the animals. This finding and previous
observations of the presence of high concentrations of
thiosulfate in the blood of field-caught clams indicate that
the animal portion of the symbiosis normally functions
aerobically. The intact symbiosis exhibited nitrate and
nitrite respiration under anoxic conditions. Nitrate res-
piration was stimulated by sulfide, as well as thiosulfate,
while nitrite respiration was only stimulated by sulfide.
Nitrate respiration also occurred when whole animals were
under oxic conditions. Respiration measurements showed
that the bacterial symbionts were capable of direct sulfide
oxidation. Sulfide-stimulated oxygen consumption by
bacterial preparations from the gills of mud-maintained
clams reached a maximum rate at 25 \iM sulfide and
showed no apparent inhibition at sulfide concentrations
up to 1 m. I/ sulfide.
Introduction
Solemya reidi is a gutless, marine protobranch bivalve
that lives in reduced sediments such as sewage outfall
Received 17 April 1 99 1 ; accepted 27 March 1992.
' Present address: Ocean Studies Board, HA-550, National Research
Council. 2101 Constitution Ave.. NW, Washington, DC 20418.
zones and pulp mill effluent sites (Reid, 1980; Reid and
Bernard, 1980). The clam contains intracellular, chemo-
lithotrophic bacterial symbionts within specialized cells
in its gills (Felbeck, 1983). The bacteria use the energy
from the oxidation of reduced sulfur compounds to fix
and reduce CO2 and subsequently translocate the fixed
carbon to symbiont-free tissues of the animal, resulting
in a net autotrophic existence of the symbiosis (Felbeck,
1983; Fisher and Childress, 1986; Anderson el ai, 1987).
This autotrophic, sulfur-dependent mode of nutrition is
called thiotrophy (Vetter, 1991).
Based on experiments done at high sulfide concentra-
tions, Powell and Somero (1985) concluded that sulfide
oxidation occurs in the animal tissue of 5. reidi and not
in the symbiotic bacteria. The authors identified intra-
cellular, ferric iron-containing granules (originally called
sulfide-oxidizing bodies, referred to as granules in this
manuscript) and based upon a colonmetric assay sug-
gested that granules were responsible for sulfide oxidation
in S. reidi (Powell and Somero, 1985). The products of
this oxidation were not identified. Subsequent studies re-
vealed that isolated mitochondria of S. reidi are also ca-
pable of sulfide oxidation and that the mitochondria cou-
ple sulfide oxidation to aerobic respiration and ATP pro-
duction (Powell and Somero, 1986: O'Brien and Vetter.
1990). The mitochondria oxidize sulfide exclusively to
thiosulfate (O'Brien and Vetter, 1990). In addition, high
concentrations of thiosulfate are found in the blood of
field-caught animals and clams experimentally exposed
to sulfide (Anderson et a/., 1987). These findings suggested
that the mitochondria may also have a role in sulfide de-
toxification.
Current models propose that the detoxification of sul-
fide is an oxygen-dependent process that occurs only in
the animal tissues of S. reidi (reviewed by Somero et ai,
1989). The oxidation results in the production of thio-
444
SOLEMYA REIDI SULFUR METABOLISM
445
sulfate. Thiosulfate is subsequently transported by the an-
imal to the bacteriocyte where it is further oxidized by
the bacteria, again by an oxygen-dependent respiratory
process.
Aspects of this model describing thiotrophic metabo-
lism are not proven or they are inaccurate because of a
lack of direct experimental evidence or the technical dif-
ficulties in early experiments. First, there is little infor-
mation on the environmental sulfide conditions in the
clam's habitat. Although the bulk sulfide concentrations
in sediment samples are highly variable, ranging from low
micromolar concentrations up to 22 mM (Childress and
Lowell, 1982; Vetter el ai, 1989), the concentration near
the animals is not known. Data suggest that the optimal
sulfide concentrations for the animals are 100 ^M or below
(Anderson el ai. 1987, 1990). In addition, the concen-
trations of other possible substrates (including thiosulfate.
which is found in the blood of freshly caught clams) in
the animals' burrows are not known.
Second, bacterial symbionts may be important agents
of direct sulfide oxidation in the symbiosis in the presence
of environmental sulfide concentrations. Recent experi-
ments in our laboratory have shown that enriched bac-
terial suspensions show stimulation of I4CO: uptake in
the presence of 500 fj.M sulfide and exhibit nitrate and
nitrite respiration under anoxic conditions in the presence
of 200 nM sulfide (Dr. Barbara Javor, pers. comm.). In
addition, in vivo measurements of spectral changes of in-
tracellular hemoglobins (Doeller el ai, 1988) and in vitro
accumulations of elemental sulfur in isolated gill ctenidia
exposed to oxygen and sulfide (Vetter. 1990) suggest that
sulfide and oxygen enter the bacteriocyte directly from
the seawater.
Third, the effect of anoxic conditions on sulfur metab-
olism of the host and symbionts needs additional study.
Anderson et ai (1990) have shown that animal tissues of
S. reidi maintain aerobic metabolism in the presence of
sulfide concentrations up to 100 ^M, but they switch to
anaerobic metabolism (fermentation) in the presence of
oxygen at higher sulfide concentrations. This switch is
due to poisoning of cytochrome c oxidase and aerobic
respiration (Anderson et ai. 1990). Whole animal exper-
iments under conditions of low O2 or at sulfide concen-
trations above 250 pAl showed a loss of net autotrophy.
suggesting that bacterial metabolism was inhibited (An-
derson el ai. 1987). However, the anaerobic sulfide-oxi-
dizing capacity of the symbionts was never measured and
the clams were incubated in surface seawater that does
not typically contain the alternate electron acceptor nitrate
that some sulfur-oxidizing bacteria can use for anaerobic
respiration. Thus, the absence of net autotrophy observed
could be due to the absence of nitrate. Net CO: uptake
might be enhanced at high sulfide concentrations or at
low O2 concentrations under conditions promoting an-
aerobic respiration in the bacteria.
High nitrate reductase activity has been observed in S.
reidi (Felbeck et ai. 1983). However, this enzyme was
proposed to be involved only in assimilatory nitrate re-
duction. The bacterial symbionts of the clam Lucinomn
aequizonata might use an anaerobic metabolic strategy
(Gary el ai, 1989), and recently nitrate respiration has
been demonstrated in the symbionts of L. aequi:onata
(U. Hentschel, pers. comm.) and S. reidi (Javor, pers.
comm.). The sulfur metabolism of the whole clam under
oxic and anoxic conditions in the presence of nitrate needs
to be investigated.
This investigation was designed to address the following
questions about sulfide and thiosulfate metabolism in the
5. reidi symbiosis. (1) Is thiosulfate present in the mud
around the clam and what causes thiosulfate to accu-
mulate in clam blood? (2) Can the bacteria oxidize sulfide
directly, and do bacteria and granules compete for sulfide?
(3) Can the intact symbiosis use nitrate and nitrite as al-
ternate electron acceptors?
Materials and Methods
Experimental animals
Specimens of Solemya reidi were collected at the Hy-
perion sewage outfall in Santa Monica Bay, California, at
a depth of 100 to 140 m using a modified Van Veen grab.
The animals were maintained in mud from their habitat
in a flowing seawater aquarium. The aquarium was kept
dark at 8°C to mimic natural conditions. All experimental
incubations were done with animals maintained less than
30 days, with most animals maintained less than 14 days.
Clam size varied from 32 to 42 mm. An effort was made
to use similar size animals within experiments. Nitrate
and nitrite respiration experiments (Figs. 1, 2) have all
values normalized to a wet gill weight of 420 mg/clam.
Blood values (Tables I, II, III) are not normalized to size
or gill weight. The values represent the concentrations
measured.
Whole clam sulfur use experiments (oxic and anoxic)
were done with aquarium-maintained animals that had
been pre-incubated for a minimum of 48 h in oxygenated
seawater (dark at 8°C) without sulfur substrates. This pre-
incubation allowed the animals to remove reduced sulfur
compounds from their blood (see Results). All incubations
were performed in a glass, water-jacketed incubation
chamber with 500 ml of filtered seawater at 10°C. Whole
clam nitrate and nitrite respiration experiments were done
with aquarium-maintained animals immediately after re-
moval from the mud. Incubations were done in ground-
glass stoppered bottles with 100 ml filtered seawater at
10°C. All anoxic incubations were done in seawater
sparged with argon for 10 min. The chambers were sealed
446
D. B. WILMOT AND R. D. VETTER
except for sampling, and argon was blown over the open-
ing during sampling.
Sulfur compound sampling and analysis
Concentrations of sulfide, thiosulfate. sulfite, and glu-
tathione were measured by HPLC using the monobro-
mobimane (bimane) technique (Newton and Fahey. 1987;
Vetter ct a/.. 1989). For porewater samples, the sediment
surrounding individual animals was collected and im-
mediately placed in 50-ml test tubes and centrifuged in a
clinical centrifuge at approximately 2500 RPM. Aliquots
of 100 |tl were immediately derivatized with bimane. The
reaction of bimane with the sample rapidly derivatized
all reduced thiols and prevented further oxidation. Blood
samples were collected by carefully cutting the membrane
that connects the two valves and opening the animal. Sea-
water was wiped away, and the mantle was cut at the
exhalent opening. Aliquots of 100 n\ blood were imme-
diately derivatized with bimane. Water samples from
whole animal incubations were collected and 100 /ul ali-
quots were immediately derivatized with bimane. Fluo-
rescence of the bimane-derivatized samples was measured
as previously described (Wilmot and Vetter. 1990).
Bacterial enrichments
Suspensions of bacterial symbionts were obtained by
gently homogenizing the gills of aquarium-maintained
clams in a glass homogenizer in filtered (0.45 tiM) sea-
water buffered with 10 m.U MOPS (3-[N-morpho-
lino]propanesulfonic acid), pH 7.5. The homogcnate was
centrifuged at low speed (53 X g) at 4°C for 5 min, pel-
leting the large cellular debris. Filtration of the suspension
through 15-/um or 28-^m Nitex filters separated the bac-
teria (approximately 1 ^m diameter) from the larger sub-
cellular particles. The filtrate was centrifuged a second
time at 1925 X ,? at 4°C for 5 min, which pelleted the
bacteria and granules while leaving the mitochondria in
the supernatant. The pellet was washed and resuspended
in MOPS-buffered seawater. The Percoll gradient tech-
nique (Distel and Felbeck, 1988; Wilmot and Vetter,
1990) was not effective in separating the bacteria from
other subcellular particles. Light microscopy was used to
confirm that preparations did not contain nuclei and large
cellular debris. Because mitochondria are too small to
detect by casual observation, the bacterial preparation was
tested for contamination by measuring ATP production.
There was no external ATP produced, indicating that ei-
ther no mitochondria were present or they could not
function in seawater. Bacterial preparations had a final
concentration of 36-38 mg wet gill tissue/ml seawater.
Bacterial enrichments were also made from clams that
had been starved for 2 1 days. These animals were placed
in oxygenated seawater containing no sulfur substrates or
nitrate and kept at 8°C in the dark. The seawater was
changed at least once a day. The bacterial suspensions
were prepared exactly as described above.
Protein determination
Total protein of bacterial suspensions was measured
by the method of Hartree (1972).
Nitrite determination
Seawater samples ( 1 ml) were collected and nitrite de-
termined using a modification of the colorimetric method
of Strickland and Parsons (1977). Due to interference by
thiols, several modifications were necessary. Briefly, we
had to solve the problem of interference by sulfide and
thiosulfate. This was accomplished by precipitating the
thiols.
Respiration measurements
All incubations were carried out in a Strathkelvin res-
piration chamber at 15°C. The respiration chamber was
thermally jacketed and contained a stir bar that allowed
homogeneous incubation at a constant temperature. Ox-
ygen consumption was measured with a Clark-type oxygen
electrode (Strathkelvin Instruments) modified to reduce
interference by H:S (O'Brien and Vetter, 1990).
Results
Sediment porewater and blood sulfur compounds
To determine which reduced sulfur compounds were
available to the clams, we sampled the sediment porewater
surrounding individual clams and in the blood of these
clams. The major sulfur compound in sediment porewater
was sulfide (Table I). Extremely low concentrations of
thiosulfate were measured, but no sulfite or glutathione
was present. Mean sulfide concentrations were highly
variable between sediment samples at different sites
(ranging from a low of 17 ^M to a high of 888 nM).
Variability within sediment samples was considerably less.
Although sulfide was the major reduced sulfur com-
pound in the seawater surrounding the animals and thio-
sulfate concentrations were extremely low, clam blood
contained high concentrations of thiosulfate (Table I).
The blood of the animals also contained low concentra-
tions of sulfide and sulfite (<12 nM). The cellular thiol
compound, glutathione, remained relatively constant
throughout sampling and was used as an indicator that
hemolymph samples were not diluted by seawater. Nitrite
was not present in either the sediment porewater or blood
from freshly collected animals (data not shown).
To determine how fast thiosulfate and sulfide could be
cleared from the blood of freshly collected animals from
SOLEMYA REIDI SULFUR METABOLISM
447
Table I
Siillur compounds from sediment rxirewalcr and blood of freshly collected Solemya reidi
Sample
Sulfur compounds (n
Sullide
Thiosulfatc
Sulfite
June, 1987, data previously published (Vetter el al
"ND" represents value not determined.
1989). Values are mean ± standard deviation.
Glutathione
Pore water
May 1987
6
385
±321
11 ±
19
ND
ND
June 1987
6
17
± 5
1 ±
0
ND
ND
JuK 1990
6
38
± 30
13 ±
->
0
0
September 1990#1
4
20
± 8
3±
1
0
0
September 1990 #2
6
888
± 60
5±
4
0
0
Blood
May 1987
6
4
± 11
297 ±
120
ND
ND
June 1987
6
1
± 0
111 ±
31
ND
ND
JuK 1990
7
11
± 6
29 +
17
<1
74 ± 27
September 1990#1
4
8
± 3
149 ±
35
9±5
28 ± 13
September 1990 #2
5
12
± 3
232 ±
89
9 ± 4
57 ±21
a single grab. 20 clams were placed in oxic seawater con-
taining no reduced sulfur substrates. The removal of thiols
from the blood of the clams is shown in Table II. An
initial thiosulfate concentration of 274 ± 106 nAl de-
creased to 17 ± 1 1 iu.\f in 12 h and to zero in 24 h. Low
concentrations of sulfide and sulfite were also cleared from
the blood within 12 h while glutathione concentrations
remained relatively constant.
Aerobic and anaerobic metabolism of sulfide and
thiosulfate by whole animals
Sulfide and thiosulfate uptake and metabolism under
oxic conditions were investigated using aquarium-main-
tained clams. Incubations with whole clams were for either
2 or 4 h. The animals were incubated in 500 ml filtered
seawater at 10°C in the presence of either sulfide or thio-
sulfate.
Table II
C Icurance <>/ sulfur compounds from the hlond <'l lrc\hl\ collected
Solemya reidi during an incubation in sulfide-free,
< i \\wnated seawater
Sulfur compounds (n.W)
Sample
n
Sulfide
Thiosulfate
Sulfite
Glutathione
T
= Oh
4
9
f
1
274 ±
106
13 ±
1
19
± 24
T
= 6h
4
17
•
8
66 ±
44
27 ±
10
27
±29
T
= 12 h
4
0
.
0
17 ±
11
0±
0
9
± 0
T
= 24 h
4
0
+
0
0±
0
2 ±
3
1 1
+ 2
T
= 96h
4
0
+
0
0±
0
0 ±
0
18
± 5
Values are mean ± standard deviation.
Whole clams incubated with sulfide (100 pM) showed
very little sulfide or sulfite accumulation in the blood dur-
ing the 2-h experiment (Table III). However, high con-
centrations of thiosulfate in the blood were observed. To
examine the possibility that thiosulfate was actively trans-
ported by the clam, a thiosulfate ratio representing the
ratio of the seawater thiosulfate concentration at the end
of each incubation to the thiosulfate concentration in the
blood at the end of each incubation was calculated. The
thiosulfate ratio under oxic conditions was 1:300. Glu-
tathione concentrations did not show any consistent pat-
tern.
Similar experiments were performed by incubating
clams with thiosulfate. Under these conditions, the clams
did not concentrate thiosulfate in their blood above the
concentration in the surrounding seawater during 2- and
4-h incubations with two thiosulfate concentrations (50
and 250 n.M) (Table III). The thiosulfate ratio for either
concentration was never greater than 1:1. Virtually no
sulfide or sulfite was observed in the blood and glutathione
again showed no consistent patterns.
We also investigated whole animal sulfide and thio-
sulfate metabolism under anoxic conditions. In these ex-
periments, whole clams were incubated in seawater that
was sparged with argon. Anoxic conditions were main-
tained throughout the incubations.
Anoxic incubations of clams with sulfide (100 /uM)
showed little sulfide or sulfite accumulation in the blood
(Table III). However, unlike oxic incubations, blood thio-
sulfate concentrations did not increase substantially (24
nM as compared to 300 nM during anoxic incubations).
Thiosulfate was removed from the seawater and ap-
peared in the blood of animals incubated under anoxic
448
D. B. WILMOT AND R. D. VETTER
Table III
Siillur cuni^iiiinils from blood o/ Solemya reidi after o.\ic and anoxic incubations with sulfidc anil llu
All incubations were in 500 ml scawaier at 10°C
Incubation conditions
Sulfur compounds (/iA/)
Sultide
Tbiosulfate
Sulfitc
Glutathione
Thiosulfate ratio*
Sulfide
Oxic— 2 h
Control
4
0
+
0
18 ±
1
0
±0
13
± 10
[00 nM
6
1
+
1
300 ±
217
1
± 1
15
+ 5
1:300
Anoxic — 2 h
Control
4
8
+
3
8±
5
4
± 3
9
± 7
100 M^
6
6
+
2
24 ±
23
7
± 1
33
± 19
1:24
Thiosulfate
Oxic— 2 h
Control
3
2
•
2
5 ±
3
4
± 1
9
± 10
250 tiM
3
4
+
1
136 ±
0
7
± 3
38
± 9
1:0.8
Oxic— 4 h
250 nM
3
3
-t-
->
152 ±
4
6
±4
38
± 2
1:0.9
Oxic— 2 h
Control
4
2
+
T
22 +
13
12
± 4
28
±20
50 M.|/
4
3
.
5
35 ±
11
9
±9
16
± 4
1:0.8
250 M I/
4
0
+
0
157 ±
17
3
+ 2
1 1
± 1
1:0.8
Anoxic — 2 h
Control
3
2
+
->
5 ±
3
4
± 1
9
± 10
250 MA/ with 5 mA/ nitrate
3
6
+
->
158±
63
4
± 1
14
± 12
1:0.7
250 MA/ w/out 5 mA/ nitrate
3
3
+
2
116 ±
33
4
+ 2
25
± 19
1:0.5
Sulfur compound values are mean ± standard deviation.
* The thiosulfate ratio represents the ratio of the seawater concentration at the end of each incubation to the concentration in the blood. The
concentration of thiosulfate in the seawater at the end of the sullide incubations was below the limits of detection, thus a value of 1 ^A/w'as assigned.
Controls represent animals after oxic or anoxic incubations without a reduced sulfur substrate (either sullide or thiosulfate).
conditions both with and without nitrate (Table III). The
pattern of blood thiols under anoxic conditions was similar
to that for oxic conditions. Thiosulfate was not concen-
trated in the blood above the concentrations of the sur-
rounding seawater under any of the incubation conditions.
Sulfidc and thiosulfate stimulated nitrate/nitrite
respiration
The intact symbiosis exhibited nitrate respiration under
oxic and anoxic conditions in the presence of sulfide and
thiosulfate (Fig. la, b). In the presence of 5 mM nitrate,
nitrite accumulated in filtered seawater during 3-h oxic
incubations with whole clams in the presence of 150 /uA/
sulfide and 250 ^M thiosulfate (Fig. la). Two different
controls were run to confirm that nitrate respiration was
being carried out by the symbiotic bacteria. First, 5 m M
nitrate was added to filtered seawater containing 1 50 ^.M
sulfide without clams. No nitrite accumulated in the sea-
water (not shown). Second, 5 mM nitrate was added to
filtered seawater containing clams and no sulfide (Fig.
la). Blood nitrite concentrations at the end of the 3-h
incubations were 158 + 80 nM (n = 4) and 52 ± 16
(n = 4) for sulfide- and thiosulfate-incubated clams, re-
spectively.
Anoxic incubations were also performed. Whole clams
had the potential for nitrate respiration under anoxic
conditions (Fig. Ib). Three different controls were run.
First, 5 mM nitrate was added to filtered, deoxygenated
seawater containing 150 n\l sulfide without clams. No
nitrite accumulated in the seawater (Fig. Ib). Second, no
nitrate was added to filtered, deoxygenated seawater con-
taining 250 n.M thiosulfate with clams. No nitrite accu-
mulated in the seawater (data not shown). Finally, 5 mM
nitrate was added to filtered, deoxygenated seawater con-
taining 150 pM sulfide with clams that had their gills re-
moved. Very little nitrite (2.0 ± 0.2 nM) accumulated in
the seawater (Fig. Ib). Blood nitrite concentrations at the
end of 3-h incubations were 214 ± 103 /uA/(n = 4) and
189 ± 80 n.M(n = 4) for sulfide- and thiosulfate-incubated
clams, respectively.
Five m.M nitrate concentrations are not environmen-
tally realistic. Because nitrite accumulation (versus nitrate
disappearance) was measured and nitrite only accumu-
lated in the presence of excess nitrate, it was essential that
excess nitrate be available to the clams during the entire
SOU-'MV ( REID1 SULFUR METABOLISM
449
A) 100
o
o
75
r so
Sulfide
Thiosulfate
Animal Control
30
60 90 120
Time (minutes)
150
125
100
O
o
Water Control
Animal Control
Sulfide
Thiosulfate
60 90 120
Time (minutes)
150
180
Figure I. Nitrate respiration of whole Solcmya rcii/i during 3-h in-
cubations in oxic and anoxic buffered seawater at 10°C. (A) Nitrite ac-
cumulation in oxic seawater during incubations containing 5 m.\I nitrate
and either 1 50 pM sulfide (mean ± standard deviation of 4 incubations)
or 250 nM thiosulfate (mean ± standard deviation of 4 incubations).
The animal control represents seawater containing 5 m,W nitrate, animals,
and no sulftde. (B) Nitrite accumulation in anoxic seawater during in-
cubations containing 5 mM nitrate and either 1 50 nM sulfide (mean
± standard deviation of 4 incubations) or 250 n\l thiosulfate (mean
± standard deviation of 4 incubations). The water control represents
seawater containing 5 mM nitrate, 1 50 fiM sulfide. and no animals. The
animal control represents seawater containing 5 mM nitrate, 150 nM
sulfide, and 4 animals with their gills removed. Clam size ranged from
32 to 42 mm and all values are normalized to a wet gill weight of 420
mg/clam.
3-h incubation. However, whole animal nitrate respiration
(nitrite accumulation) rates similar to those presented
above were found when clams were incubated in 500 n\l
and 50 n.\I nitrate for short incubations (only single ex-
periment— data not shown). Similar results have also been
found for bacterial preparations (Javor. pers. comm.).
The intact symbiosis showed nitrite respiration under
anoxic conditions in the presence of sulfide but not thio-
sulfate (Fig. 2). In the presence of 150 nM sulfide, 100
n.M nitrite was removed from filtered seawater by whole
clams incubated anoxicallv. An initial nitrite concentra-
tion of 10 1 .5 ± 2.8 nM decreased to 69.9 ± 6.0 ^I/during
a 3-h incubation. Two different controls were run to con-
firm that nitrite respiration, like nitrate respiration, was
being carried out by the symbiotic bacteria. First, lOOjtA/
nitrite was added to filtered, deoxygenated seawater con-
taining 1 50 pM sulfide and no clams. The seawater nitrite
concentration did not decrease (Fig. 2). Second, nitrite
was added to filtered, deoxygenated seawater containing
no reduced sulfur compound and three clams. Again, the
seawater nitrite concentration did not decrease (data not
shown). Blood nitrite concentrations at the end of 3-h
incubations were 5.3 ± 2.5 n\l and 17.3 ± 4.6 n\l for
sulfide and thiosulfate incubations, respectively.
Sulfide-stimulated oxygen consumption by bacterial
symbionts
The whole animal experiments could not directly ad-
dress the question of whether the bacteria could carry out
respiratory sulfide oxidation. To determine if the sym-
bionts could respire sulfide aerobically, we measured
azide-sensitive (respiratory) and azide-insensitive (non-
respiratory) oxygen consumption of bacterial suspensions
in the presence of different sulfide concentrations. The
bacterial suspensions contained granules because they
could not be separated from the bacteria. The suspensions
oxidized a wide range of sulfide concentrations with max-
imal respiration rates at 25 nM sulfide (Fig. 3a). A pair-
wise comparison of values at 25 fiM and all higher con-
centrations using the Mann-Whitney U test (Zar, 1984)
found no significant difference between 25 pM and any
higher values (P > 0.05). However, the number of repli-
100
80
c
o
6°
40
01
Z 20
Water Control
Sulfide
Thiosulfate
30 60 90 120
Time (minutes)
150
180
Figure 2. Nitrite respiration of whole Solcmya reidi during 3-h in-
cubations in anoxic buffered seawater at !0°C. Nitrite disappearance in
seawater during incubations containing either 200 ^M sulfide (mean
± standard deviation of 4 incubations) or 200 pM thiosulfate (mean
± standard deviation of 4 incubations). The water control represents
seawater containing 1 00 M^/ nitrite, 1 50 jjA/ sulfide, and no clams. Clam
size ranged from 32 to 42 mm and all values are normalized to a wet
gill weight of 420 mg/clam.
450
D. B. WILMOT AND R. D. VETTER
A) 20
200 400 600
Sulfide Concentration
1000
B)
0 200 400 600 800 1000
Sulfide Concentration (nM)
Figure 3. Sultide-stimulated oxygen consumption in buffered sea-
water at 15°C by bacterial enrichments from (A) the gills of mud-main-
tained Solemya rcidi or (B) the gills of clams maintained in oxygenated
seawater that contained no reduced sulfur compounds or nitrate for 21
days. Data points represent means and standard deviations from three
experiments (n = 3). See Materials and Methods for clam size and bac-
terial density.
cates (n = 3) is not sufficient for a robust statistical de-
termination. Thus, aerobic respiration was apparently not
inhibited by sulfide concentrations up to 1 ni/U. However,
the azide-insensitive oxygen consumption rate increased
over the range of sulfide concentrations tested. The max-
imal rate for azide-insensitive consumption was above 1
mM The data in Figure 3a are the combination of the
means from multiple measurements in three different ex-
periments (each experiment was done with one bacterial
preparation consisting of many clams).
Similar respiration experiments were conducted on
bacterial suspensions from clams that were maintained
in oxygenated seawater without an external reduced sulfur
source or nitrate for three weeks. The gills of these animals
were dark (no yellow sulfur globules). The total number
of bacteria or a bacteria to granule ratio was not deter-
mined. However, it has been observed that starved clams
have more granules than healthy clams (pers. obs.). The
bacterial suspensions from these gills showed a different
pattern for azide-sensitive and insensitive oxidation rates
(Fig. 3b). Azide-sensitive aerobic respiration again had a
maximum rate at 25 ^M, but appeared to be inhibited at
higher concentrations with complete inhibition at 500 n\l
and greater sulfide concentrations. The azide-insensitive
rate increased with increasing sulfide concentrations and
was more than double that found in the bacterial enrich-
ments from animals freshly removed from mud based on
total protein of the suspensions. The data in Figure 3b
are the combination of three different experiments.
Discussion
Analysis of sediment porewaters showed that sulfide
was typically the only reduced sulfur compound available
to S. reicli and that concentrations ranged from below 20
nM to nearly 1 m.\/. Because the clams actively pump
water from above the sediment through their burrows,
sulfide concentrations in the water that is in contact with
the gills may be significantly lower. Porewater thiosulfate
was present in extremely low concentrations, and the
clams were unable to accumulate thiosulfate in the blood
above the external concentration. Thus, it appears that
porewater thiosulfate is of little importance as an energy
source in the intact symbiosis.
Similar to Anderson ct al. (1987), we observed high
concentrations of thiosulfate in the blood of field caught
animals and animals experimentally incubated in the
presence of sulfide. Our data show that thiosulfate present
in the blood of freshly collected clams results from host
oxidation of sulfide and not from uptake of porewater
thiosulfate. Because thiosulfate accumulates in the blood
during oxic incubations with sulfide but not during anoxic
incubations, molecular oxygen must be required for sul-
fide oxidation to thiosulfate. It is most likely that aerobic
respiration of sulfide by mitochondria is responsible for
thiosulfate production (O'Brien and Vetter. 1990). Thio-
sulfate is cleared from the blood by the bacteria within
bacteriocytes. However, the relative importance of this
energy source versus direct uptake of sulfide across the
gill is not known.
Two clearly different types of sulfide oxidation occurred
in our bacterial enrichments. The first, true respiration,
represents electron transport chain (ETS)-linked bacterial
sulfide oxidation and it is completely azide-sensitive. The
ETS-linked type showed a high capacity to oxidize sulfide
at low concentrations (maximal rate by 25 nM). The sec-
ond, which is azide-insensitive, represents non-enzymatic
oxidation by ferric iron catalysis (hematin) in granules
(Powell and Arp, 1989) and autocatalysis by sulfur (Chen
and Morris, 1972).
Non-enzymatic, heat-stable catalysis of sulfide oxida-
tion has been observed in a variety of animal tissues (re-
SOU-MYA RE1D1 SULFUR METABOLISM
451
viewed by Beauchamp et ai. 1984) and specifically in a
thiotrophic symbiosis (Wilmot and Vetter. 1990). It has
been shown that ferric iron-containing compounds such
as hematin and ferritin are responsible for the non-en-
zymatic catalysis in mammals (Sorbo. 1958: Baxter and
van Reen, 1958; Baxter et al.. 1958). The benzyl viologen
(BV) assay used to determine which components of the
5. rciili symbiosis carried out sulfide oxidation (Powell
and Somero, 1985) measures non-enzymatic ferric iron
catalysis in the presence of high ( 1-5 m.U) sulfide (Powell
and Arp, 1989). It does not measure ETS-linked activity
because high sulfide concentrations inhibit the oxidation.
Thus, it is not surprising that early studies using the BV
assay concluded that the bacteria and mitochondria did
not oxidize sulfide (Powell and Somero, 1985).
Although we have shown that the bacteria are capable
of sulfide oxidation, it is not clear whether the bacteria
encounter sulfide or if it is first oxidized by mitochondria
or electron-dense granules. Several lines of evidence sug-
gest that the bacteria normally oxidize sulfide as well as
thiosulfate in vivo. First, the bacteria are oriented close to
the outside and are separated from the seawater by a thin
epithelial cell that contains few mitochondria or electron-
dense granules (Felbeck, 1 983; Gustafson and Reid, 1988).
Because the bacteria are not packed close to the blood
space, which contains the mitochondria! product of sulfide
oxidation, it does not seem likely that thiosulfate is the
only sulfur substrate available to the bacteria. Second,
isolated gill ctenidia can oxidize sulfide without a host
blood supply (Vetter. 1990). The ctenidia produce ele-
mental sulfur and protein, which seems to indicate that
sulfide and oxygen are taken directly across the gill (Vetter.
1990). Third, the bacteriocytes of S. reidi and S. velum
contain intracellular hemoglobin that can bind sulfide as
ferric hemoglobin sulfide (Doeller et al.. 1988; Krause
and Wittenberg, 1990). Fiber-optic spectroscopy of intact
gills indicated that ferric hemoglobin sulfide was present
in the bacteriocyte region of the gill.
The data presented above do not provide direct proof
that the bacteria within a bacteriocyte oxidize sulfide. Ja-
vor (pers. comm.) has recently observed that bacterial
suspensions from S. reidi respire nitrate under oxic and
anoxic conditions in the presence of sulfide and thiosul-
fate. Similarly, in this study, the symbionts respired nitrate
that entered the bacteriocyte, either from the blood, or
directly from seawater in the presence of sulfide and thio-
sulfate. The product of the anaerobic respiration, nitrite,
was excreted from the bacteriocyte in the presence of ex-
cess nitrate and accumulated in the blood and seawater.
In marine denitrifying bacteria, nitrate is often reduced
only to nitrite (Goering and Cline. 1970) and the nitrite
accumulates outside the cells (Payne and Riley, 1969). If
nitrate concentrations become low or are exhausted, the
nitrite can be taken back up by the bacteria and further
reduced.
In the absence of nitrate, the intact symbiosis respired
nitrite in the presence of sulfide only. This has also been
observed for bacterial suspensions from S. reidi (Javor,
pers. comm.). The nitrite, like the nitrate, either enters
the bacteriocyte from the blood or directly from seawater.
More importantly, these data provide the best available
evidence that the bacteria within bacteriocytes oxidize
sulfide. These results suggest that sulfide oxidation is cou-
pled to complete denitrification, but that thiosulfate ox-
idation is coupled to the first step only. Anaerobic sulfide
oxidation by the bacteria may be an important detoxifi-
cation mechanism when the clam is depleted of oxygen.
In addition, because aerobic nitrate respiration has been
described for several bacteria (Robertson and Kuenen.
1984; Lloyd et ai. 1987) denitrification may occur when
nitrate (and sulfide) is present and oxygen concentrations
are low.
Although the symbionts are capable of denitrification,
and the bacteria have access to nitrate and nitrite when
each is present in seawater, we do not know the concen-
trations that are available to the animals in the natural
environment. The water flow across the sludge field is
strong and the oxygen concentrations and water chemistry
are similar to the surrounding areas (approximately 100-
175 nMO2)(B. Thompson, pers. comm.). At neighboring
hydrographic stations in the Santa Monica Bay area, the
ammonia concentrations range from 0.5 to 2.0 ^M. while
sediment porewaters typically are 100-1000-fold higher
(Eppley. 1986). Nitrifying bacteria are active in these wa-
ters (Ward et al.. 1 982) and nitrate values at a depth equal
to the sludge field are >20 n.\l (Williams, 1986). Nitrate
is not present below the top 2 cm of sediment at Whites
Point outfall near the Hyperion outfall (J. Gieskes, pers.
comm.). It can be assumed that as the clams pump water
from above the sediment through their burrows, they are
probably exposed to oxygen and nitrate simultaneously,
but the ratio of oxygen to nitrate is unknown.
When the data presented here are integrated with pre-
vious studies (Felbeck, 1983; Powell and Somero, 1985,
1986; Fisher and Childress. 1986: Anderson et al.. 1987,
1990; Doeller et al.. 1988; Krause and Wittenberg, 1990;
O'Brien and Vetter. 1990). a more complete picture
emerges of how the symbiosis may be functioning under
different environmental conditions. Maximum net au-
totrophy occurs in the presence of external sulfide con-
centrations of 100 n.W (Anderson et al.. 1987). Because
seawater is near pH 8. 1 at 100 fiAt total sulfide, approx-
imately 3 n.W sulfide as H:S diffuses into animal cells
(Millero. 1986). Assuming an internal pH near pH 7.5,
the maximal internal free sulfide (H2S) concentration is
approximately 20 nAf, which is very close to the aerobic
sulfide-oxidizing maximum for both isolated mitochon-
452
D. B. WILMOT AND R. D. VETTER
dria (O'Brien and Vetter, 1990) and bacteria. Aerobic
metabolism is maintained by animal tissues in the pres-
ence of external sulfide concentrations up to 100 fiM. At
higher concentrations, the onset of anaerobic pathways is
evident (Anderson el ai. 1990).
Sulftde oxidation can occur by at least two processes:
sulfide respiration by mitochondria and bacteria, and non-
enzymatic catalysis by metal containing granules, and au-
tocatalysis by sulfur. The mitochondria oxidize sulfide
only to thiosulfate which accumulates in the clam's blood
(O'Brien and Vetter, 1990). The oxidation is oxygen-de-
pendent and can yield ATP (Powell and Somero, 1986).
Presumably, the bacteria oxidize sulfide to elemental sul-
fur and polysulfides and ultimately to sulfate. Bacterial
sulfide respiration can occur aerobically when oxygen is
available or possibly anaerobically when nitrate (or nitrite)
is available as an alternate electron acceptor. The impor-
tance of nitrate (nitrite) respiration is unknown. In ad-
dition, the thiosulfate produced by the mitochondria can
be further oxidized by the bacteria. The importance of
the non-enzymatic oxidation of sulfide is unknown.
We have shown the metabolic potential for the use of
nitrate and nitrite as alternate electron acceptors in the
intact symbiosis. Future studies must determine the en-
vironmental concentrations of nitrate, nitrite, and oxygen
and the relevance of anaerobic respiration to the sym-
biosis. If oxidized nitrogen compounds are present, they
are likely in the oxygen-containing water drawn into the
burrow by the clams. Thus, conditions that limit oxygen
availability would also limit nitrate. A reasonable hy-
pothesis may be that a switch to anaerobic nitrogen res-
piration by the symbionts is a tactic by the symbiosis to
save oxygen for host metabolism during times of low ox-
ygen.
Acknowledgments
We gratefully acknowledge the captains and crews of
the research vessels R1' Robert Gordon Sproul and R\
Point Sur. especially Louis Zimm, without whom this
work would not have been possible. We would like to
thank Dr. George Somero for reviewing the manuscript,
the laboratories of H. Felbeck and J. Childress for assisting
on cruises and providing collecting opportunities, and
John O'Brien for his many hours of assistance at sea and
in the laboratory. Special thanks is due Ron Kaufmann
for helpful discussions regarding data analysis and Dr.
Barbara Javor for reviewing the manuscript and for her
laboratory assistance and stimulating discussions during
these studies. This work was supported by NSF grant
OCE86-10513 (to G. N. Somero and R. D. Vetter) and
ONR grant NOOO 14-87-00 12 (to R. D. Vetter).
Literature Cited
Anderson, A. E., J. J. Childress, and J. Favuzzi. 1987. Net uptake of
CO: driven by sulfide and thiosulfate oxidation in the bacterial sym-
biont-containing clam Soleniyu reidi. J. Exp. Biol. 133: 1-31.
Anderson, A. E., H. Felbeck, and J. J. Childress. 1990. Aerobic me-
tabolism is maintained in animal tissue during rapid sulfide oxidation
in the symbiont-containing clam Solemya reidi. J. Exp. Zoo/. 256:
130-134.
Baxter, C. F., and R. van Reen. 1958. Some aspects of sulfide oxidation
by rat-liver preparations. Biochim. Biophys. Ada 28: 567-573.
Baxter, C. F., R. van Recn, P. B. Pearson, and C. Rosenberg.
1958. Sulfide oxidation in rat tissues. Biochim. Biophys. Ac/a 27:
584-591.
Beauchamp, R. O., Jr., J. S. Bus, J. A. Popp, C. J. Boreiko. and D. A.
Andjelkouch. 198-4. A critical review of the literature on hydrogen
sulfide toxicity. Cm. Rev Toxicol. 13( 1 ): 25-97.
Cary, S. C., R. D. Vetter, and H. Felbeck. 1989. Habitat characterization
and nutritional strategies of the endosymbiont-beanng bivalve Lu-
ciiwma aequi-onala. Mar Ecu/ Prog. Ser 55: 31-45.
Chen, K. V., and J. C. Morris. 1972. Kinetics of oxidation of aqueous
sulfide by O:. Emir. Sci Techno/ 6(6): 529-537.
Childress, J. J., and \V. Lowell. 1982. The abundance of a sulftde-
oxidizing symbiosis (the clam Solemya reidi) in relation to interstitial
water chemistry. Am /.ool. 63: A45.
Distel, D. L., and H. Felbeck. 1988. Pathways of inorganic carbon
fixation in the endosymbiotic-beanng lucinid clam Liicinoma ae-
1/111 :i»ntta. Part 1. Purification and characterization of the endosym-
biotic bacteria. J. Exp. Bio/. 247: 1-10.
Doeller, J. E., D. \V. Kraus, J. M. Colacino, and J. B. Wittenberg.
1988. Gill hemoglobin may deliver sulfide to bacterial symbionts
ofSo/emya velum (Bivalvia. Mollusca). Biol. Bull 175: 388-396.
Fpplev, R. \V. 1986. People and the plankton. Pp. 289-303 in Lecture
Notes on Coastal and Esniarine Studies. Vol. 1 5, Plankton Dynamic^
ol the Southern California Bight. R. W. Eppley, ed. Springer- Verlag.
New York.
Kelbeck, H. 1983. Sulfide oxidation and carbon fixation by the gutless
clam Solemya reuli an animal-bacterial symbiosis. J Comp. Physiol.
152: 3-11.
Fisher, C. R.. and J. J. Childress. 1986. Translocation of fixed carbon
from symbiotic bacteria to host tissue in the gutless bivalve. Solemya
reidi. Mar Biol 93: 59-68.
Goering, J. J., and J. C. Cline. 1970. A note on denitrification in sea-
water. Limnnl. Oceanogr. 15: 306-308.
Gustafson. R. G., and R. G. B. Reid. 1988. Association of bacteria with
larvae of the gutless protohranch bivalve Solemya reidi. Mar Biol.
97: 389-401.
Hartree, II. F. 1972. Determination of protein: a modification of the
Lowry method that gives a linear photometric response. Anal.
Biochem 48: 422-427.
Kraus. D. \\ '., and J. B. \\ittenberg. 1990. Hemoglobins of the Lucina
pectinata/bacteria symbiosis I. Molecular properties, kinetics and
equilibria of reactions with ligands. J. Biol. Chem 265: 16043-16053.
Lloyd, D., L. Boddy, and J. P. Davies. 1987. Persistence of bacterial
denitrification capacity under aerobic conditions: the rule rather than
the exception. FEMS Microbiol. Ecol. 45: 185-190.
Millero, F.J. 1986. The thermodynamics and kinetics of the hydrogen
sulfide system in natural waters. Mar. Chem 18: 121-147.
Newton, G. L., and R. C. Fahey. 1987. Purification of thiols from bio-
logical samples. Melh Eniymol. 143:96-100.
O'Brien, .]., and R. D. Vetter. 1990. Production of thiosulfate during
sulphide oxidation by mitochondria of the symbiont-containing bi-
va\\e Solemyu reidi ./. Exp Biol 149: 133-148.
Payne, \V. J., and P. S. Riley. 1969. Suppression by nitrate of enzymatic
reduction of nitric oxide. Proc. Soc. Exp. Biol Med. 132: 258-260.
SOIJiMYA RE1D1 SULFUR METABOLISM
453
Powell, M. A., and G. N. Somero. 1985. Sullidc oxidation occurs in
the animal tissue of the gutless clam, Solemya reicii. Bio/ Bull 169:
164-181.
Powell, M. A., and G. N. Somero. 1986. Hydrogen sulfide oxidation
is coupled to oxidative phosphor, lation in mitochondna ofSolemya
rcidi. Science 233: 563-566.
Powell, M. A., and A. J. Arp. 1989. Hydrogen sulfide oxidation by
abundant nonhemoglobin heme compounds in marine invertebrates
from sulfide-rich habitats. J. E.\p. Zoo/. 249: 121-132.
Reid, R. G. B. 1 980. Aspects of the biology of a gutless species of So-
Icniya (Bivalvia: Protobranchia) Can. J. Zoo/. 58: 386-393.
Reid, R. G. B.. and F. R. Bernard. 1980. Gutless bivalves. Science 208:
609-610.
Robertson, L. A., and J. G. Kuenen. 1984. Aerobic denitrification: a
controversy revived. Arch. Microbiol. 139: 351-354.
Somero, G. N., J. J. Childress, and A. K. Anderson. 1989. Transport,
metabolism and detoxification of hydrogen sulfide in animals from
sulfide-rich marine environments. Crii. Rev A(/nal. Sci. 1: 591-
614.
Sorbo. B. 1958. On the oxidation of thiosulfate from inorganic sulfide
by rat tissue and heme compounds. Biochnn. Biophys. Ada 27: 324-
329.
Strickland. J. D. H.. and T. R. Parsons. 1977. A practical handbook
of seawater analysis, 2nd ed. Bull. Fish. Res Board Can. 167: 1 -
310.
Vetter, R. D. 1990. Cultured gill filaments from Solemya reidi: a model
system for the study of thiotrophic symbioses. Pp. 349-352 in En-
docytnbiology /I '. P. Nardon, V. Gianinazzi-Pearson. A. M. Grenier,
L. Margulis, and D. C. Smith, eds. Institute National de le Recherche
Agronoique. Paris. France.
Vetter, R. D. 1991. Symbiosis and the evolution of novel trophic strat-
egies: thiotrophic organisms at hydrothermal vents. Pp. 219-245 in
Symbiosis as a Source of Evolutionary Innovation: Special/on and
Morphogenesis, L. Margulis and R. Fester, eds. MIT Press, Cam-
bridge, Massachusetts.
Vetter, R. D., P. A. Matrai, B. Javor, and J. O'Brien. 1989. Reduced
sulfur compounds in the marine environment: analysis by HPLC.
Pp. 244-261 in Biogenic Sulfur in the Environment. E. S. Sa'.tzman.
and W. J. Cooper, ed. American Chemical Society, Washington, DC.
Ward, B. B., R. J. Olson, and M. J. Perry. 1982. Microbial nitrification
rates in the primary nitrite maxima oft" Southern California. Deep-
Sea Res. 29: 247-255.
Williams. P. W. 1986. Chemistry' of the dissolved and paniculate phases
in the water column. Pp. 53-83 in Lecture Notes on Coasta/ and
Estuarine Studies I'ol I?. Plankton Dynamics oj the Southern Cal-
ifornia Bight. R. W. Eppley, ed. Sponger- Verlag, New York.
Wilmot, Jr., D. B., and R. D. Vetter. 1990. The bacterial symbiont
from the hydrothermal vent tubeworm Rijtia pachyplila is a sulfide
specialist. Mar. Biol 106: 273-283.
Zar, J. H. 1984. Biostatistical Analysis, second edition. Prentice-Hall.
Inc. Englewood Cliffs, New Jersey.
Reference: Binl Bull 182: 454-457. (June. 1992)
To Thine Own Self be True? An Addendum to
Feldgarden and Yund's Report on Fusion
and the Evolution of Allorecognition in
Colonial Marine Invertebrates
RICHARD K. GROSBERG1
Department of Zoology, Center for Population Biology.
University of California, Davis, California 95616
Feldgarden and Yund ( 1 ) recently re-examined a ques-
tion central to understanding the evolution of allorecog-
nition systems in colonial and clonal marine invertebrates:
"How does the genetic polymorphism necessary to restrict
intergenotypic fusion to kin and clonemates accumulate
in natural populations?" They argue that explanations
invoking kin selection account fully for neither the ex-
tensive allelic polymorphism that characterizes the genetic
systems that control allorecognition specificity, nor the
apparent phenotypic instability of genetic chimeras. As a
more parsimonious explanation than kin selection for
these observations, they propose that frequency-dependent
selection, acting at the level of the individual, promotes
the accumulation of allotypic polymorphism by favoring
fusion with self (and preventing fusion with nonself). To
place this proposition into a conceptual framework, they
cite two of my papers (2, 3) and assert that, "Several au-
thors have noted that kin selection does not provide an
obvious explanation for high allotype diversity." As far
as it goes, this is an accurate statement, at least with respect
to intergenotypic fusion in populations lacking kin struc-
ture. In these papers, however, J. F. Quinn and I developed
a series of analytical models showing how individual se-
lection could easily maintain allotypic polymorphism
through the control of intergenotypic fusion.
So that there can be no misunderstanding about what
we said, I quote a summary of our work on allotypic spec-
ificity and fusion from a paper published in 1988 (2, pp.
402-403; full references can be found in the original text):
Received 10 February 1992; accepted 10 March 1992.
It appears intuitively that individual selection acting to re-
strict allogeneic fusion could represent a general and potent
selective force favoring the evolution of allorecognition and
allotypic specificity (Burnet, 1971, 1973; Buss, 1982). In a
theoretical analysis of this problem. Grosberg and Quinn
(1988) defined the conditions necessary to favor rare al-
lot) pic variants in a single-locus, haploid model. Let c, be
the net per-capita fitness cost of fusion and b, be the net
fitness gain that is due to fusion. The expected fitness of
an allorecognition allele, /, upon which fusion is condi-
tioned, is then
W, =
-- c/)
(12)
where P, is the frequency of allele /. Equation (12) shows
that the fitness of an allorecognition allele depends upon
both its frequency and the relative costs and benefits of
fusion. If b, is greater than c/, then W, will increase as P,
increases and the allele will become fixed in the population.
However, if c, is greater than b,, then as P, increases, W,
decreases. Consequently, rare alleles will be favored and
allotypic polymorphism will accumulate only when the
costs effusion exceed the benefits. This raises the question
of why individuals should ever fuse (Grosberg and Quinn,
1988)'
One of the important effects of high levels of allotypic
variation is the restriction of fusion to closely related in-
dividuals. Although the costs and benefits of genotype fu-
sion should be adjusted according to the relatedness of the
fused colonies (Hamilton. 1964; Buss and Green, 1985;
Grosberg and Quinn. 1986), the effects of kin selection
have not yet been incorporated into mathematical models
of allotypic specificity.
I believe this passage leaves little room for misinter-
pretation, either of our approach to the problem, or of
454
TO THINE OWN SELF BE TRUE?
455
the result that frequency-dependent selection, acting at
the level of the individual, can favor rare alleles and pro-
mote the evolution of polymorphism at allorecognition
loci through the restriction of intergenotypic fusion. Nev-
ertheless, a number of biological complexities should
temper the conclusion that individual selection is the pri-
mary force maintaining allotypic variation. First, as Feld-
garden and Vund ( 1) mention, the theoretical prediction
that individual selection can maintain allotypic poly-
morphism does not preclude the operation of kin selec-
tion. In fact, kin selection may be particularly effective in
the many taxa of sedentary, clonal invertebrates in which
sibling sexual propagules tend to remain spatially asso-
ciated via restricted dispersal (4-8) or preferential settle-
ment near kin (6, 9). Such a pattern of larval dispersal
will, even after only a single generation, lead to kin struc-
ture, increased probabilities effusion, and the opportunity
for kin selection. Moreover, there is ample evidence that
sexually produced propagules of some sponges, cnidarians,
and ascidians can and do fuse soon after settlement (2).
Thus, their claim [based on (6)] that. "Larval settlement
as a function of future fusibility is the sole observation
that we are aware of that is consistent with kin selection
but not with the selective pressure of self fusion," could
well apply to many clonal marine invertebrates.
Second, if the preservation of clonal integrity is the
primary function of allorecognition systems, why should
fusion ever be permitted between non-clonemates? In ge-
netic terms, the overriding problem is: why, in many taxa,
is only partial, rather than complete, genetic matching
among alleles at allorecognition loci required for fusion
to occur (2, 10)? After all. allorecognition systems re-
quiring only partial allotypic matching, and based on self-
recognition, are far less efficient at prohibiting fusion with
nonself than systems requiring complete matching (in
which any allelic discrepancy in allotype would lead to
rejection and the preservation of genotypic integrity) (11).
One answer [the one favored by Feldgarden and Yund
(1)] is that intergenotypic fusion is simply a matter of
imprecision in the recognition system and may be of little
or no selective importance: "Although fusion between kin
occurs, such events may simply represent mistakes in rec-
ognition due to the limitation of an imperfect system." It
could be. as Feldgarden and Yund contend, that the ge-
netics of invertebrate allorecognition are biochemically
and phylogenetically constrained, so that partial genetic
matching, and the recognition errors that it entails, is an
evolutionary necessity, whatever the selective optimum.
I think, however, that this response oversimplifies even
the meager amount presently known about the cellular
and genetic mechanisms that regulate allorecognition. In
so doing, it begs several crucial observations pertinent to
the evolution of allorecognition and intergenotypic fusion.
In particular, if there were no exceptions to the genetic
rule of partial matching in the clades that Feldgarden and
Yund cite, then their explanation retains substantial merit.
However, in several of the invertebrate phyla that they
mention, including some sponges, cnidarians, and ascid-
ians, full allotypic matching appears to be required for
fusion to occur (12-16). Similarly, recognition systems
based on multiple independent loci are less prone to error
than single locus systems with comparable levels of allelic
variation (11). Although some well-studied taxa (e.g.. the
compound ascidian Bolnilus) have primary allorecog-
nition systems based on a single locus, other taxa appear
to have allotypic markers encoded by several loci (17, 18).
Finally, in the few systems that have been examined in
any sort of detail, individual genotypes appear to distin-
guish among different classes of nonself (2, 10, 19). I do
not know why there is such variation, but taken together,
these observations suggest, at least, that more precise al-
lorecognition systems can evolve, but often do not.
There are four other poorly characterized, but none-
theless crucial, aspects of allorecognition that further
complicate our understanding of how allotypic specificity
evolves. First, although some [but not all, £'.,?., (20)] genetic
chimeras have been found to be morphologically unstable
(21-23), little is known of the genetic stability of these
chimeras (24) and thus the true costs and benefits of fu-
sion. Moreover, in the colonial ascidian Botnilus xtiilos-
.vt'/v, the morphological stability of chimeras seems to de-
pend on the relatedness. and perhaps allotypic similarity,
effusion partners ( 19-23).
Second, in the absence of clone-specific genetic probes,
the frequency of chimera formation in natural populations
of benthic invertebrates is notoriously difficult to estimate.
In general, grafting assays imply that intergenotypic fusion
should be rare, provided that there is little kin structure
in a population (2). Thus, I am not surprised that taxa
such as Hydraciinin symbiologicarpus, which inhabit
mobile substrata, spawn gametes, or have motile, free-
swimming larvae, show little evidence of kin structure
and consequently little evidence for natural chimeras (25).
In contrast, other sessile species that live on fixed surfaces
and brood low vagility. sexual offspring, ought to have
much higher frequencies of intergenotypic fusion. This
might be the case for other hydractiniids, such as Hy-
dractinia mi/ltri. and is known to be the case in Botryllm
sch/osseri (7).
Third, as Feldgarden and Yund acknowledge, it is es-
sential to quantify the costs and benefits of fusion, and
how these might vary with ontogenetic, genetic, and eco-
logical circumstances. If the situation in a chimera is as
simple as one genotype's fitness loss being exactly the oth-
er's fitness gain, then, over the long term, it is difficult to
see whv even the most closelv related nonclonemates
456
R. K. GROSBERG
should be allowed to fuse (except by recognition error).
On the other hand, if costs and benefits depend on on-
togenetic, genetic, or ecological factors, or if costs and
benefits are not additive, then kin selection may be effec-
tive.
Finally, Feldgardcn and Yund did not consider contexts
other than intergenotypic fusion in which allotypic spec-
ificity regulates the nature and outcomes of interactions
between conspecifics. For example, in many cnidarians.
allotypic disparity leads to aggressive behavior, whereas
similarity fails to elicit an aggressive response (2). For this
set of alternative behaviors, both Crozier (26) and Gros-
berg and Quinn (3) showed that individual selection does
not provide a straightforward explanation for the mainte-
nance of allotypic variation; with kin structure, however,
polymorphism can evolve (27). In still more complex sit-
uations, pure fusion or aggression models are unrealistic.
For instance, in Hydractinia symbiolongicarpus, incom-
patible colonies usually behave aggressively, whereas
compatible genotypes often somatically fuse (28). Theo-
retical analysis of these behavioral options predicts that
allotypic variation can be maintained, but only if fusion
is more costly than aggression (3).
The paper by Feldgarden and Yund does focus atten-
tion on the idea that the preservation of clonal integrity
(which is an extreme form of kin selection) can be an
important selective mechanism, an idea first articulated
nearly a century ago by Bancroft (29), and echoed over a
quarter century ago by Knight-Jones and Moyse (30) and
Hamilton (31). The paper further helps to clarify how
little we know about the genetics and fitness consequences
of allorecognition and intergenotypic fusion. Until more
of this sort of information is in hand for a variety of taxa,
we should not consider recognition errors and their effects
on inclusive fitness as being mere epiphenomena of im-
perfect allorecognition systems. Consequently. I am re-
luctant— even in the face of having shown how individual
selection can maintain allotypic specificity in the context
of fusion — to downplay the potential importance of kin
selection.
Acknowledgments
Thanks to D. Brumbaugh. B. Johnson, and C. Pfister
for their helpful comments. I am also grateful to A. T.
Hun and R. T. Paine for their insightful suggestions. The
National Science Foundation has supported this research.
Literature Cited
1. Feldgarden, M., and P. O. Yund. 1992. Allorecognition in colonial
marine invertebrates: does selection favor fusion with kin, or fusion
with self? fi;,-/ Bull 182: 155-158.
2. Grosberg, R. k. 1988. The evolution of allorecognition specificity
in clonal invertebrates. Q. Rev Rial 63: 377-412.
3. Grosberg, R. K.. and J. F. Quinn. 1988. The evolution of allorec-
ognition specificity. Pp. 157-167 in Invertebrate Historecognition,
R. K. Grosberg, D. Hedgecock, and K. Nelson, eds. Plenum Press.
New York.
4. Jackson, J. B. C. 1985. Distribution and ecology of clonal and
aclonal benthic invertebrates. Pp. 297-356 in Population Biology
and Evolution »l C/onal Organisms, J. B. C. Jackson. L. W. Buss,
and R. E. Cook, eds. Yale University Press. New Haven.
5. Jackson, J. B. C. 1986. Modes of dispersal of clonal benthic in-
vertebrates: consequences for species' distributions and genetic
structure of local populations. Bull Mar. Sei. 39: 588-606.
6. Grosberg, R. K., and J. F. Quinn. 1986. The genetic control and
consequences of kin recognition by the larvae of a colonial ascidian.
Nature 322: 456-459.
7. Grosberg, R. K. 1987. Limited dispersal and proximity-dependent
mating success in the colonial ascidian Botryllm xchloxaeri. Evolution
41: 372-384.
8. Keough, M. J., and H. Chernoff. 1987. Dispersal and population
variation in the bryozoan Biigula nerilina. Ecology 68: 199-210.
9. Keough, M. J. 198-1. Kin-recognition and the spatial distribution
of larvae of the bryozoan Bngiila nerilina (L.). Evolution 38:
142-147.
10. Neigel, J. E. 1988. Recognition of self or nonself? Theoretical
implications and an empirical test. Pp. 127-142 in Invertebrate His-
toreeogiiilion. R. K. Grosberg. D. Hedgecock. and K. Nelson, eds.
Plenum Press. New Y'ork.
1 1. Curtis, A. S. G., J. Kerr, and N. Knotvlton. 1982. Graft rejection
in sponges. Genetic structure of accepting and rejecting populations.
Tran^i'/iinluiion 33: 127-133.
12. Neigel, J. E., and J. C. Avise. 1983. Histocompatibility bioassays
of population structure in marine sponges. ./. Heredity 14: 134-140.
1 3. Neigel, J. E., and J. C. Avise. 1983. Clonal diversity and population
structure in a reef-building coral, Acropora cervicornis: self-recog-
nition analysis and demographic interpretation. Evolution 37: 437-
453.
14. Neigel, J. E., and J. C. Avise. 1985. The precision of histocom-
patibility response in clonal recognition in tropical marine sponges.
Evolution 39: 724-732.
15. \Vulff, J. L. 1986. Variation in clone structure of fragmenting coral
reel sponges. Biol ./ Linn Sin 27: 31 1-330.
16. Raftos, D. A., D. A. Briscoe, and N. N. Tail. 1988. The mode of
recognition of allogeneic tissue in the solitary urochordate Styelu
I'lieata. Transplantation 45: 1123-1126.
17. Fuke, M. T., and I. Nakamura. 1985. Pattern of cellular alloreac-
ti\it\ of the solitary ascidian, Halocynthia roret:i. in relation to ge-
netic control. Biol. Bull 169: 631-637.
18. Hauenschild, C. von. 1956. Uber die Vererbung einer Gewebev-
ertraglichkeits-Eigenschaft bei dem Hydroidpolypen Hydractinia
et'liiiuilii Z .\iitnr/or\cli 116: 132-138.
19. Weissman, I. L., V. L. Scofield, V. Saito, H. Boyd, and B. Rinkevich.
1988. Speculations on the relationships of two Botryllm allorecog-
nition reactions — colony specificity and resorption — to vertebrate
histocompatibility. Pp. 67-78 in Invertebrate HiMoreeoi;nilioii.
R. K. Grosberg. D. Hedgecock, and K. Nelson, eds. Plenum Press,
New Y'ork.
20. Ilan, M., and Y. Loya. 1990. Ontogenetic variation in sponge his-
tocompatibility responses. Biol. Bull. 179: 279-286.
21. Rinkevich, B., and 1. L. \\eissman. 1987. A long-term study on
fused subclones in the ascidian Bolryllin, schlowri: the resorption
phenomenon (Protochordata: Tunicatal. ./. Zool. (LoiulJ 213: 717-
733.
22. Rinkevich, B., and I. 1.. Weissman. 1987. The fate of Botryllu*
larvae cosettled with parental colonies: beneficial or deleterious con-
sequences? Biol. Bull 173:474-488.
TO THINE OWN SELF BE TRUE?
457
2V Rinketii-h. B.. and I. L. \\eissman. 1989. Variations in Ihe out-
comes following chimera formation in the colonial tunicate Boiryllus
vY/A.v.vcn. Bull. Mar Sci. 45: 213-227.
2-4. kingsley. E. \., D. \. Briscoe, and D. A. Raftos. 1989. Correlation
of histocompatibility reactions with fusion between conspeeincs in
the solitary urochordate Styela plicata Biol Hull 176:282-289.
25. Yund, P. O.. and H. M. Parker. 1989. Population structure of
Hydractinia sp. nov. C in the Gulf of Maine. ./. /;.v/> Mar. Bio. Ec'ol.
125:63-82.
26. Crozier, R. H. 1986. Genetic clonal recognition abilities must
be retained by selection for something else. Evolution 40: 1100-
27. Ratnieks, F. I.. \\ . 1991. Evolution of discriminatory aggression
in marine invertebrates. ./ Theor. Bio/. 152: 557-565.
28. Buss, L. W., C. S. McFadden, and D. R. Keene. 1984. Biology of
hydractiniid hydroids. 2. Histocompatibility effector system/com-
petitive mechanisms mediated by nematocyst discharge. Bio/ Bull
167: 139-158.
29. Bancroft, F. \V. 1903. Variation and fusion ot colonies in com-
pound ascidians. Proc Calil. Aanl. Sci. (Series 3) 3: 137-186.
30. Knight-Jones, K. \V.,and J. Mnyse. 1961. I ntraspecific competition
in sedentary marine animals. Symp. Soc. Exp. Biol. 15: 72-95.
3 1 . Hamilton, \V. D. 1964. The genetical evolution of social behaviour.
/ Theor. Biol. 7: 1-52.
Reference: Biol. Bull 182: 458-459. (June, 1992)
To Thine Own Self be True? Yes! Thou Canst Not
Then be False to Any Other. A Reply to Grosberg
PHILIP O. YUND1 AND MICHAEL FELDGARDEN2
Department oj Biological Sciences, University of New Orleans, New Orleans, Louisiana 70148 and
2 Department of Biology, Yale University, New Haven, Connecticut 06511
Our original paper ( 1 ) was motivated by the philosophy
that every healthy scientific discipline should have at least
two alternative hypotheses (2); our goal was simply to
incorporate a consideration of selection favoring self fu-
sion in future work on the selective forces operating on
allorecognition systems. We did not intend to resolve this
issue, and we do not feel that a resolution is possible given
the limitations of current information. Consequently, we
welcome Grosberg's commentary (3) on our previous ef-
forts and are thankful for this opportunity to extend the
discussion. Grosberg raises a number of interesting points
about our argument, three of which merit additional
commentary.
First, Grosberg feels that we have misinterpreted a series
of models developed with J. F. Quinn (4). We acknowledge
that our discussion of their results was too brief to fully
justify our interpretation. Nevertheless, we do feel that our
interpretation is valid. Grosberg and Quinn did consider the
costs and benefits of fusion and aggression at the individual
level. However, the costs and benefits were explicitly those
associated with fusion with kin. Neither fusion with self, nor
any other selective force, was considered as an alternative.
Our contribution was an explicit consideration of the selec-
tive forces of self fusion, which can be invoked only at the
individual level. We did not intend to imply that past work
had ignored all forms of selection at the individual level,
just the selective pressures of fusion with self.
An additional aspect of our interpretation of Grosberg
and Quinn's results involves their predictions from a series
of models that consider the consequences of variation in
costs and benefits for three cases (4): aggression alone,
fusion alone, and fusion as an alternative to aggression.
The aggression only model predicts a monomorphic pop-
ulation under all conditions. In both of the remaining
cases, polymorphism occurs only when the costs exceed
Received 23 March 1992; accepted 26 March 1992.
the benefits [as cited in Grosberg's quote (3)]. When the
benefits are greater than the cost, the initially most fre-
quent allele increases to fixation, and the population be-
comes monomorphic. We interpret this result as prelim-
inary evidence against fusion with kin generating and
maintaining high levels of polymorphism. The scenario
of costs exceeding benefits is compatible with neither the
basic assumptions behind the evolution of kin fusion, nor
the continued existence of allorecognition systems. While
Grosberg may now feel that these models show that "in-
dividual selection could easily maintain allotypic poly-
morphism" (3, p. 454), he and Quinn originally reached
rather a different conclusion that is in accord with our
own interpretation of their work:
"allotypic polymorphism can be maintained directly by
the individual costs and benefits effusion provided fusion
carries a net fitness cost. This raises the question of how
fusion conditioned on relatedness can be evolutionary
stable. Our results suggest that selection acting at the level
of clonal or kin-aggregations, rather than at the level of the
individual, may provide an explanation for the evolution of
allotypic specificity through aggression or fusion." (4, p. 157)
"Because the individual costs and benefits of fusion and
aggression cannot readily account for why these behaviors
are conditioned on allotypic identity, other explanations
must be sought." (4, p. 165)
While Grosberg and Quinn's models do predict poly-
morphism under some conditions, polymorphism occurs
only under a much more restrictive set of circumstances
than predicted by selection for fusion with self. A model
incorporating both organismal and genotype level effects
of kin fusion might lead to a different conclusion, as
Grosberg and Quinn suggest (4).
As a second point of discussion, Grosberg mentions ev-
idence of the widespread occurrence of fusion soon after
larval settlement as support for the potentially broad impact
of kin fusion. We do not dispute that many colonial taxa
458
\ RHPLY TO GROSBERG
459
have the capability to fuse soon after metamorphosis.
However, the ability of individuals to fuse or reject upon
assuming a benthic existence is equally compatible with
the selective pressures of both self and kin fusion. Only the
occurrence of aggregated larval settlement based on future
fusibility is incompatible with the self fusion hypothesis.
To the best of our knowledge, evidence of aggregated set-
tlement as a function of shared allorecognition alleles is
available only for the ascidian Bolryllm schlosseri. Although
larvae of the bryozoan Biigula neriiina preferentially settle
near relatives (5), subsequent allorecognition interactions
between colonies in this species are not known.
Third, Grosberg questions why an error-prone single
locus system, based on partial match rules, would evolve
in botrylloid ascidians if selection for fusion with self is
occurring. He points out that other taxa, including other
ascidians. possess multiple locus and full match rule al-
lorecognition systems that would generate fewer errors in
recognizing and fusing with self. This is certainly a valid
argument. While we feel that phylogenetic constraints are
a real possibility that should be seriously considered as
an alternative to adaptationist explanations (6). we also
recognize the possible validity of Grosberg's interpretation.
We have already acknowledged that allorecognition sys-
tems in botrylloid ascidians show effects of kin selection
due to the existence of aggregated settlement of fusible
larvae ( 1 ). This does not imply that the same is necessarily
true of all other taxa. We chose botrylloids as one of our
examples because they are one of the very few taxa for
which there is enough information to conduct a prelim-
inary evaluation of both hypotheses, not because we felt
that they provided the strongest support for our idea;
clearly they do not.
However, there is a second logical conclusion that can
be drawn from Grosberg's line of reasoning about genetic
mechanisms of allorecognition. Single locus, partial match
control of allorecognition in botrylloid ascidians may
generate a relatively high frequency of fusion events be-
tween kin. By the same logic, multiple locus, full match
systems are likely to result in an exceedingly low frequency
of fusion among kin in other taxa. For example, a two
locus system with full match rules and high levels of poly-
morphism at both loci (i.e., such that most individuals
are heterozygous and parents do not share alleles) will
result in all rejection responses between offspring and
parents and a fusion frequency among randomly paired
full siblings of only 1/16. Multiple paternity of broods
will further reduce the fusion frequency among siblings,
as half sibs will not fuse. Linkage between allorecognition
loci will increase the rate of fusion among sibs, but only
to a maximum frequency of 1/4 (in the case of a zero
recombination rate, functionally equivalent to a single
locus system). These low fusion rates among close relatives
will greatly reduce the potential for kin selection to impact
allorecognition systems in other taxa. Only if aggregated
settlement based on future fusibility is prevalent will kin
fusion be common in colonial taxa with multiple locus,
full match rules. A reduction in the frequency of fusion
with kin is likely to increase the relative frequency of fu-
sion with self (i.e., proportion of fusion events that occur
with self vs. kin), increasing the potential for selection for
self fusion to influence the system.
We feel it very unlikely that either of these two hy-
potheses can be definitively excluded on the basis of cur-
rent information on allorecognition systems. More em-
pirical work is clearly required. The task is especially dif-
ficult because the two selective forces are not mutually
exclusive, and their relative impact is likely to vary among
taxa. Although it has traditionally been very difficult to
distinguish between self and kin fusion events (7-9), the
application of current molecular techniques should render
these problems much more tractable. We do not advocate
that the possibility of kin selection be abandoned, just
that future empirical work consider alternative selective
forces as well.
Acknowledgments
We thank John Francis, Steve Gaines, Mike Mc-
Cartney, and Pam O'Neil for their comments and John
Francis for supplying the full text of Polonius' speech to
Laertes. Funding was provided by the National Science
Foundation and the Louisiana Stimulus for Excellence in
Research (NSF/LaSER( 1991 )-RCD-03).
Literature Cited
1. Feldgarden, M., and P. (). Yund. 1992. Allorecognition in colonial
marine invertebrates: does selection favor fusion with kin, or fusion
with self? Biol. Bull. 182: 155-158.
2. Platt, J. R. 1964. Strong inference. Science 146: 347-353.
3. Grosberg, R. K. 1992. To thine own self be true? An addendum
to Feldgarden and Yund's report on fusion and the evolution of
allorecognition in colonial marine invertebrates. Biol. Bull. 182: 454-
457.
4. Grosberg, R. K., and J. F. Quinn. 1988. The evolution of allorec-
ognition specificity. Pp. 157-167 in Invertebrate Historecognition.
R. K. Grosberg. D. Hedgecock. and K. Nelson, eds. Plenum Press.
New York.
5. Keough, M. J. 1984. Kin-recognition and the spatial distribution
of larvae of the brvozoan Biigulu neriiina (L.). Evolution 38: 142-
147.
6. Gould, S.J., and R.C.Lewontin. 1979. The spandrels of San Marco
and the Panglossian paradigm: a critique of the adaptationist pro-
gramme. Proc R Sue Loud. B205: 581-598.
7. Curtis, A. S. G., J. Kerr, and N. Knowlton. 1982. Graft rejection
in sponges: genetic structure of accepting and rejecting populations.
Transplantation 30: 362-367.
8. Neigel, J. E., and G. P. Schmahl. 1984. Phenotypic variation within
histocompatibility-defined clones of marine sponges. Science 224:
413-415.
9. Neigel, J. E., and J. C. Arise. 1985. The precision of histocom-
patibility response in clonal recognition in tropical marine sponges.
Evolution 39: 724-732.
INDEX
Ahalone lysin cDNA sequences, 97
Abalone sperm lysin. 97
Acoela, 54
Adult plasticity and rapid larval evolution in a recently isolated barnacle
population. 210
Age of the mangrove crab Scylla serraia at colonization by stalked bar-
nacles of the genus Oclolasmis, 1 88
Alcyonaria. 231
ALFORD, NICOLE, see Christian A. Combs, 416
Allogeneic fusion, 155
Allorecognition, 155. 221. 454. 458
Allorecognition in colonial marine invertebrates: does selection favor
fusion with kin, or fusion with self? 155
AMEMIYA, S., AND R. B. EMLET, The development and larval form of
an echinothurioid echinoid, AMhcnn\,>nia i/imai. 15
Amphioxus, 11
Anemonia viriilis. 159
Annulate lamellae. 41
Anoxia, 265
Antipatharia, 195
Antipredator defensed in tropical Pacific soft corals (Coelenterata: Al-
cyonacea). I. Sclentes as defense against generalist carnivorous fishes.
231
. tp/y\ui. 8
Architectural and mechanical properties of the black coral skeleton
(Coelenterata: Antipatharia): a comparison of two species. 195
Are temperature and photoperiod necessary' cues for encystment in the
marine benthic harpacticoid copepod Heteropsyllus mtnni Coull?
109
Ascidians, 458
Asexual reproduction, 169
Asteroidea. 177
ATP-reactivated models. 382
Ausio, JUAN, Purification and biochemical characterization of the nuclear
sperm-specific proteins of the bivalve mollusks Agriodesma saxicnla
and Mytilimeria iniitalli, 31
Autogeneic fusion. 155
\\oidance of hypoxia in cnidarian symbiosis by algal photosynthetic
oxygen. 159
B
Bacterial symbiosis. 105
BAGHDASARIAN, GAREN, see Ruth D. Gates, 324
Balamfi amphitrote, 210
BAKER. S., AND R. MANN, Effects of hypoxia and anoxia on larval set-
tlement juvenile growth, and juvenile survival of the oyster Cras-
sostrca virginica, 265
Becarcoidea lulandn. 4 Id
Behavior. 401
Behavioral osmoregulation, 416
Behavioral regulation of hemolymph osmolarity through selective
drinking in land crabs. Birs>u.\ latru and (jccamridea lalandii. 416
BERGLES, DWIGHT. AND SIDNEY TAMM, Control of cilia in the branchial
basket of dona insiestinalis (Ascidacea), 382
Bioluminescence. 391
Biomechanics. 424
Bivalve mollusks. 31
Bivalve reproduction, 145
Black coral skeleton. 145
BLADES-ECKELBARGER, PAMELA I.. AND NANCY H. MARCUS. The Ongin
of cortical visicles and their role in egg envelope formation in the
"spiny" eggs of a calanoid copepod. Centropages vclificaiiu*. 41
BLOCK, GENE D., see Nancy L. Wayne. 8
BOYNTON. ANGELA, see Christian A. Combs. 416
Branchial basket, 382
Brooding, 177
BROWN. CHRISTINE A.. AND NORA B. TERWILLIGER. Developmental
changes in ionic and osmotic regulation in the Dungeness crab,
Cancer inagister, 270
Bryozoans. 22 1
Burgus lain i. 416
BURTON, RONALD S., Proline synthesis during osmotic stress in megalopa
stage larvae of the blue crab, Catlmectes sapidus, 409
Calanoida. 41
Calcium control, 382
Callinectes sapidua, 409
Cancer magisit'i; 270
CANNON. LESTER B., see Joseph B. Jennings. 1 17
Capitellid, 129
CASE. JAMES F.. see Michael A. Latz. 391
Causes and consequences of fluctuations coelomic pressure in sea urchins.
424
Cccinpia. 165
Cell culture, 66
Cell detachment. 324
Cenlrnpagt.'*. 41
CHANDLER. RESA M., MARY BETH THOMAS. AND JULIAN P. S. SMITH,
III. The role of shell granules and accessory cells in eggshell formation
in ( 'nm'ii/itia pulchru (Turbellana, Acoela), 54
Characterization of two novel neuropeptides from the sea cucumber
Holoihuna glabcrrima, 241
Chemical defense. 105
CHIA. FU-SHIANG, RON Koss, SHAUNA STEVENS. AND JEFF I. GOLD-
BERG. Isolation of neurons of a nudibranch veliger. 66
Chitin. 145
Cilia, control, reversal. 248
Ciliary arrest, 382
Ciliary inactivation, 382
Ciona. 382
Circannual life cycle. 289
Clione limacina, 1
Clonal invertebrates. 454
Cnidarian symbiosis. 159
Cmdanans, 324
COLACINO. JAMES P.. see Charlotte P. Mangum. 124
Collagen, 117
Colonial invertebrates. 155. 221. 458
COMBS. CHRISTIAN A.. NICOLE ALFORD. ANGELA BOYNTON. MARK
DVORNAK, AND RAYMOND P. HENRY, Behavioral regulation of
hemolymph osmolarity through selective drinking in land crabs,
Birgits Intro and Gecarcoidea lalandii. 416
460
INDEX TO VOLUME 182
461
Compound eye, 278
Control of cilia in the branchial basket of Ciona instestinalis (Ascidacca),
382
Convo/uW eggshell formation, 54
COOK, JOHN S.. see Dorothy M. Skinner, 165
COON. S. L.. see William K. Fitt. 401
Copepoda. 41, 109
Coral reefs, 231
Corpora allata. 1 64
Cortical granules. 41
Cortical reaction, 4 1
Com L. BRUCE C.. see Judy Williams-Howze. 109
Counterillumination, 391
Crab. 270
Crassostmi virifinica. 265
Crayfish. 333. 341
CRONIN. THOMAS W.. Visual rhythms in stomatopod crustaceans ob-
served in the pseudopupil. 278
Crustaceana. 41. 270, 333, 391
Cryptic species, 1 29
Culture. 169
Culture, sexual and asexual reproduction, and growth of the sea anemone
Nemaimiclla vMensis. The. 169
Cvcloheximide. 409
D
Defenses. 231
Development. 169, 177
Development and larval form of an echinothunoid echinoid, Astheno-
siiinu uii'uii. The, 15
Developmental changes in ionic and osmotic regulation in the Dungeness
crab. Cancer magislcr. 270
Diapause, 109
DlAZ-MlRANDA, LUCY, DAVID A. PRICE. MICHAEL J. GREENBERG.
TERRY D. LEE, KAREN E. DOBLE, AND JOSE E. GARCI'A-ARRARAS,
Characterization of two novel neuropeptides from the sea cucumber
Holothuria glaberrima, 241
DIMOCK, RONALD V'.. JR., see Richard A. Tankersley, 145
Divergence of populations, 210
Dnergence ot species-specific abalone sperm lysins is promoted by pos-
itive Darwinian selection. The, 97
DOBLE. KAREN E., see Lucy Diaz-Miranda. 241
DOUGLAS. A. E.. see M. L, Rands, 159
DVORNAK. MARK, see Christian A. Combs. 416
E
Early development in the lancelet (= amphioxus) Brachioslimui /loriiiac,
from sperm entry through pronuclear fusion: presence of vegetal
pole plasm and lack of conspicuous ooplasmic segregation, 77
Echinodermata. 177, 241
Echinoid, 15
Effects of hypoxia and anoxia on larval settlement juvenile growlh, and
juvenile survival of the oyster Crassoslrea virgmica. 265
Effects of photoperiod and temperature on egg-laying behavior in a marine
mollusk, 8
Egg envelope, 41
Egg laying. 8
Eggshell granule, 54
Electrophysiology. 167
ELLERS, OLAF. AND MALCOLM TELFORD. Causes and consequences of
fluctuations coelomic pressure in sea urchins. 424
Embryology. 305
Embryos of Hiimunm amcricamts are protected by epibiotic bacteria.
105
EMLET. R. B.. see S. Amemiya, 15
Encapsulated development, 305
Encystment. 104
Energy metabolism. 298
Epibiotic bacteria. 105
Epidermal scales. 1 1 7
Epidermis, 1 I 7
Evidence for a programmed circannual life cycle modulated by increasing
daylength in Scanlhcs limnicola (Polychaeta: Nereidae) from central
California. 289
Evidence for ammonia as a natural cue for recruitment of oyster larvae
to oyster beds in a Georgia salt marsh. 401
Evolution, 177, 210, 454
Fast-strike feeding behavior in a pteropod mollusk, Clionc liiiniciiiii
Phipps. 1
Feeding. 1
FELDGARDEN. MICHAEL, AND PHILIP O. YUND, Allorecognition in co-
lonial marine invertebrates: does selection favor fusion with kin. or
fusion with self? 155
FELDGARDEN. MICHAEL, see Philip O. Yund, 458
FENICAL, WILLIAM, see M. Sofia Gil-Turnes. 105
Fertilization. 77. 197
FINGERMAN, MlLTON, see Gunderao K. Kulkarni. 341
FITT. WILLIAM K., AND S. L. COON. Evidence for ammonia as a natural
cue for recruitment of oyster larvae to oyster beds in a Georgia salt
marsh. 401
FMRFamide. 241,333
FONG, PETER P.. AND JOHN S. PEARSE, Evidence for a programmed
circannual life cycle modulated by increasing daylength in Neanthes
limnicola (Polychaeta: Nereidae) from central California. 289
Frequency-dependent selection, 155, 454, 458
Freshwater, 145
Gamete recognition proteins, 97
GARCIA-ARRARAS, JOSE, see Lucy Diaz-Miranda. 241
Gastropods, 305
GATES, RUTH D.. GAREN BAHGDASARIAN. AND LEONARD MUSCATINE,
Temperature stress causes host cell detachment in symbiotic cni-
darians: implications for coral bleaching. 324
Giant axons and escape swimming in Euplokamis dunlapae (Ctenophora:
Cydippida). 248
GIL-TURNES. M. SOFIA. AND WILLIAM FENICAL, Embryos ofHomum\
amencanus are protected by epibiotic bacteria, 105
Gills, 145
GOLDBERG. JEFF II.. see Fu-Shiang Chia, 66
GOLDBERG, WALTER M., see Kiho Kim, 195
Goose barnacle, 188
GRASSLE, JUDITH P., see Charlotte P. Mangum, 129
GREENBERG, MICHAEL J., see Lucy Diaz-Miranda, 241
GROSBERG, RICHARD K., To thine own self be true? An addendum to
Feldgarden and Yund's report on fusion and the evolution of al-
lorecognition in colonial marine invertebrates, 454
Growth, 424
Growth hypoxia. 265
H
Haliolis, 97
Hamarus tiincncuiui^. 105
HAND. CADET. AND KEVIN UHLINGER, The culture, sexual and asexual
reproduction, and growth of the sea anemone NenialnxU'llii uwr/;w\.
169
HAROSI, F. L, see K. V. Singarajah, 135
Hemoglobin. 129
Hemolymph, 270
HENRY, RAYMOND P., see Christian A. Combs. 416
HERMANS. COLIN O.. AND RICHARD A. SATTERLIE, Fast-strike feeding
behavior in a pteropod mollusk. Clione limacina Phipps. 1
HICK. ADRIAN J., see Joseph B. Jennings, 1 1 7
Histocompatibility. 155,454
462
INDEX TO VOLUME 182
HOLLAND, LINDA Z., AND NICHOLAS D. HOLLAND, Early development
in the lancelet (= amphioxus) Brachiostama jloridae, from sperm
entr>' through pronuclear fusion: presence of vegetal pole plasm
and lack of conspicuous ooplasmic segregation. 77
HOLLAND, NICHOLAS D., see Linda Z. Holland. 77
Holothurians. 24 1
How the axon got its tale, 167
Hydrogen sultide reduction of symbiont cytochrome i\i2 m gi"s of So-
leinva reidi (Mollusca). 435
Hydroids. 458
Hydrostatic skeleton, 1
Hypoxia avoidance, 159
In-vivo3'PNMR. 159
I
Intercolony coordination of zooid behavior and a new class of pore plates
in marine bryozoan, 22 I
Intercolony interactions, 22 1
Intracellular optical physiology, 278
Invertebrate. 241, 270
Invertebrate vision, 278
Ion regulation. 270
Irrigation, 145
Isolation and dissociation of ganglia, 66
Isolation of neurons of a nudibranch veliger, 66
JEFFRIES, WILLIAM B., HAROLD K. VORIS, AND SOMBAT POOVACHIR-
ANON, Age of the mangrove crab Scyllu xerrata at colonization by
stalked barnacles of the genus Octolasmis, 188
JENNINGS, JOSEPH B.. LESTER R. G. CANNON, AND ADRIAN J. HICK.
The nature and origin of the epidermal scales of Notodactylus hund-
scluni — an unusual temnocephalid turbellarian ectosymbiotic on
crayfish from northern Queensland, 1 17
Juvenile hormone. 165
Juveniles, 265
K
KIM. KIHO, WALTER M. GOLDBERG, AND GEORGE T. TAYLOR, Ar-
chitectural and mechanical properties of the black coral skeleton
(Coelenterata: Antipatharia): a comparison of two species. 195
Kin selection, 155
KOSS. RON, see Fu-Shiang Chia, 66
KRAUS, DAVID W.. JEANETTE E. DOELLER. AND JONATHAN B. WIT-
TENBERG. Hydrogen sulfide reduction of symbiont cytochrome c'S52
in gills ol Si'/emyii rculi (Mollusca), 435
KUKARNI, GUNDERAO K., AND MlLTON FlNGERMAN, Quantitative
analysis by reverse phase high performance liquid chromatography
of 5-hydroxytryptamine in the central nervous system of the red
swamp crayfish. Procambarus clarkii. 341
Lancclel, 11
Land crabs. 4 1 d
Laplace's Laws, 424
Larva. 15. 177. 348.401
Larval characters. 210
kidney. 305
release. 257
settlement, 265
LATZ. MICHAEL A., AND JAMES F. CASE. Slow photic and chemical
induction of bioluminescence in the midwater shrimp, Sergesles
xinulix Hansen, 39 1
LEE, TERRY D., see Lucy Diaz-Miranda. 241
LEE, YouN-Ho, AND VICTOR VACQUIER. The divergence of species-
specific abalo-ie sperm lysins is promoted by positive Darwinian
selection. 97
Lipid body. 298
LOLIGHMAN. B. C., see M. L. Rands, 159
Lysin, 97
Lyicchinii.i. 424
M
MACKIE, G. O., C. E. MILLS, AND C. L. SINGLA, Giant axons and escape
swimming in Euplokamis ditnlapae (Ctenophora: Cydippida), 248
MandiKa, 165
MANGUM, CHARLOTTE P.. JAMES M. COLACINO. AND JUDITH P. GRAS-
SLE, Red blood cell oxygen binding in capitellid polychaetes. 129
MANN, R., see S. Baker, 265
MARCUS, NANCY H., see Pamela I. Blades-Eckelbarger. 41
Marine Harpacticoida, 1(19
MCEDWARD. LARRY R., Morphology and development of a unique type
of pelagic larvae in the starfish Pteraster tesselalux ( Echmodermata:
Asteroidea). 177
MELLON, DEFOREST. JR., How the axon got its tale. 167
Membrane physiology, 167
Memhnmipora membranacea, 221
MERCIER. A. JOFFRE, AND RUNE T. RUSSENES, Modulation of crayfish
hearts by FMRFamide-related peptides. 333
Metamorphosis. 401
MEYER, KAREN, see Kathryn L. Van Alstyne, 231
Micronesia, 231
Microspectrophotometry. 135
Microvilli, as templates. 1 1 7
Midwater. 391
MILLS. C. E., see G. O. Mackie, 248
MITA. MASATOSHI, AND MASARU NAKAMURA, LUtrastructural study of
an endogenous energy substrate in spermatozoa of the sea urchin.
Hemicentrotus pulcherrimus, 298
Modified development. 1 5
Modulation of crayfish hearts by FMRFamide-related peptides. 333
Mollusca. 1, 8, 66," 435
Moiphogenesis, 15
Morphology and development of a unique type of pelagic larvae in the
starfish Pieraxier tesselatus (Echinodermata: Asteroidea). 177
Morphometric analysis, 145
Mortality. 265
Mud crab, 188
MUSCATINE, LEONARD, see Ruth D. Gates. 324
N
NAKAMURA, MASARU, see Masatoshi Mita. 298
Nature and origin of the epidermal scales of Notodactylus handtchini —
an unusual temnocephalid turbellarian ectosymbiotic on crayfish
from northern Queensland, The, 1 17
\eanihes, 289
Nervous conduction. 167
Nervous system. 348
Neurodevelopment, 348. 366
Neurons. 66
Neuropeptides. 241. 333
New interpretation of a nudibranch central nervous system based on
ultrastructural analysis of neurodevelopment in Mclihc lamina. I.
Cerebral and visceral loop ganglia. 348
New interpretation of a nudibranch central nervous system based on
ultrastructural analysis of neurodevelopment in Mclihc leonina. II.
Pedal, pleural. and labial ganglia, 366
Nitrate respiration, 444
Nonfeeding larvae, 15
Nolodactylus handichiiu. 1 1 7
Nuclear sperm-specific proteins. 31
Nudihranch. 66
O
O: affinity. 12')
Octolasmis cur, 188
INDEX TO VOLUME 182
463
Ontogeny. 188, 270
Oocyte. 41, 54
Ooplasmic segregation, 7"
Opisthobranch. 348
Origin of conical visicles and their role in egg envelope formation in the
"spiny" eggs of a calanoid copepod, Centropages velificattix. The,
41
Osmoregulation, 270. 409
Osphradium. 366
Oxygen- and nitrogen-dependent sulfur metabolism in the thiotrophic
clam Solemya rcidi, 444
Ovsters. 265, 401
RANDS, M. L., A. E. DOUGLAS. B. C. LOUGHMAN, AND R. G. RATCLIFFE.
Avoidance of hypoxia in cnidarian symbiosis by algal photosynthetic
oxygen. 159
RATCLIFFE, R. G., see M. L. Rands. 159
Recruitment. 401
Red blood cell oxygen binding in capitellid polychaetes. 129
Rhabdites. 117
Rhythm. 278
RIVEST, BRIAN R., Studies on the structure and function of the larval
kidney complex of prosobranch gastropods. 305
Role of shell granules and accessory cells in eggshell formation in Con-
voluta pulchra (Turbellaria, Acoela), The, 54
RUSSENES, RUNE T., see A. Joffre Mercier, 333
PAGE, LOUISE R., New interpretation of a nudibranch central nervous
system based on ultrastructural analysis of neurodevelopment in
Melibe leonina. 1. Cerebral and visceral loop ganglia. 348
PAGE. LOUISE R.. New interpretation of a nudibranch central nervous
system based on ultrastructural analysis of neurodevelopment in
Mclibc leonina. II. Pedal, pleura), and labial ganglia. 366
PAUL, VALERIE J.. see Kathryn L. Van Alstyne, 231
PEARSE. JOHN S.. see Peter P. Fong, 289
Penstomial membrane. 424
Phase shift of a tidal rhythm by light-dark cycles in the semi-terrestrial
crab Scsarma picluin, 257
Phenotypic plasticity. 210
Photic habitat, 135
Photopenod. 8, 289
Photophore. 391
Ph\logeny, 348
Planula. 169
Pleural ganglia. 348
Pluteus. 15
Pneu. 424
Polychaete. 289
POOVACHIRANON, SoMBAT, see William B. Jeffries. 188
Poreplates, 221
Positive Darwinian selection. 97
Post-colonization. 188
Predation. 231
Pressure. 424
PRICE. DAVID A., see Lucy Diaz-Miranda. 241
Procaiiihiinn clurkn. 341
Proline synthesis during osmotic stress in megalopa stage larvae of the
blue crab. Callinecles sapidus, 409
Pronuclear fusion, 77
Pronuclear movements, 77
Protamines, 31
Protonephndia. 305
Pseudopupil, 278
Pteropod, 1
Pumping, 145
Purification and biochemical characterization of the nuclear sperm-spe-
cific proteins of the bivalve mollusks Agriodesma saxicola and A/r-
lilimeria nutlalli. 3 1
Quantitative analysis by reverse phase high performance liquid chro-
matography of 5-hydroxytryptamine in the central nervous system
of the red swamp crayfish, Procambarus clarkii. 341
Quantitative analysis of the structure and function of the marsupial gills
of the freshwater mussel Anndonia cataracla, 145
RAIMONDI, PETER T., Adult plasticity and rapid larval evolution in a
recently isolated barnacle population, 210
SAIGUSA, MASAYLIKI, Phase shift of a tidal rhythm by light-dark cycles
in the semi-terrestrial crab St'.iarma picluni, 257
Salton Sea, 210
SATTERLIE, RICHARD A., see Colin O. Hermans, 1
Scales, in a turbellarian. 1 17
Sclentes, 231
Scylla serrata, 18f
Sea anemone. 169
Sea cucumbers. 241
Sea urchin sperm, 298
Seasonal reproduction, 8
Self/nonself recognition, 454
Sensory physiology, 1 35
Sergestes, 39 1
Serotonin, 391
Settlement, 401
Sexual reproduction, 169
SHAPIRO, DANIEL F., Intercolony coordination of zooid behavior and a
new class of pore plates in marine bryozoan, 221
Shrimp, 391
SINGARAJAH, K. V., AND F. I. HAROSI. Visual cells and pigments in a
demersal fish, the black sea bass (Cenlropristis striala), 135
SINGLA. C. L., see G. O. Mackie, 248
SKINNER, DOROTHY M., AND JOHN S. COOK, CARROLL M. WILLIAMS.
165
Slow photic and chemical induction of bioluminescence in the midwater
shrimp, Sergestes simi/is Hansen, 391
SMITH, JULIAN P. S.. Ill, see Resa M. Chandler. 54
Spat. 265
Sperm entry, 77
Sperm lysin cDNA, 97
Sperm lysins. 97
Sperm-egg recognition, 97
STEVENS, SHAUNA. see Fu-Shiang Chia, 66
Stomatopod, 278
Strongylocentrotus, 424
Studies on the structure and function of the larval kidney complex of
prosobranch gastropods, 305
Sulfide oxidation, 444
Sulfur metabolism in Sulcmya mdi, 444
Sulfur-oxidizing bacteria, 444
Survival. 265
Symbiodinium, 1 59
Symbiosis. 188.324.435.444
Synapse, neuro-ciliary. 248
TAMM, SIDNEY, see Dwight Bergles, 382
TANKERSLEY, RICHARD A., RONALD V. DIMOCK. JR.. Quantitative
analysis of the structure and function of the marsupial gills of the
freshwater mussel Anodonta calaracla. 145
TAYLOR, GEORGE T., see Kiho Kim. 195
TELFORD, MALCOLM, see Olaf Ellers, 424
TemnocephalidcL, 1 1 7
464
INDEX TO VOLUME 182
Temperature, 8, 324
Temperature stress causes host cell detachment in symbiotic cnidarians:
implications for coral bleaching, 324
TERWILLIGER, NORA B.. see Christine A. Brown, 270
Thiotrophic metabolism, 444
THOMAS, MARY BETH, see Resa M. Chandler, 54
Tidal rhythm, 257
To thine own self be true? An addendum to Feldgarden and Yund's
report on fusion and the evolution of allorecognition in colonial
marine invertebrates, 454
To thine own self be true? Yes! Thou canst not then be false to any
other. A reply to Grosberg, 458
Turbellaria, 54
Turbellarian, epidermal scales in an unusual, 1 17
u
UHLINGHER, KEVIN, see Cadet Hand. 169
Ultrastructural study of an endogenous energy substrate in spermatozoa
of the sea urchin. Hcmiccntr<>tm pulchcrrnniis, 298
infrastructure, 298
Unionids. 145
Units of selection. 155. 458
Urchin. 424
Vegetal pole plasm. 77
Veliger larva. 66
VETTER, RUSSELL D., see David B. Wilmot, 444
Visual cells and pigments in a demersal fish, the black sea bass (Centro-
pn.slis striata), 135
Visual rhythms in stomatopod crustaceans observed in the pseudopupil.
278'
VORIS. HAROLD K.. see William B. Jeffries, 188
vv
WAYNE. NANCY L., AND GENE D. BLOCK, Effects of photoperiod and
temperature on egg-laying behavior in a marine mollusk, 8
WILLIAMS, CARROLL M., 165
WiLLlAMS-HowzE, JUDY, AND BRUCE C. COULL, Are temperature and
photoperiod necessary cues for encystment in the marine benthic
harpacticoid copepod HclcrnpsyHiix inmni Coull? 109
WILMOT, DAVID B., AND RUSSELL D. VETTER, Oxygen- and nitrogen-
dependent sulfur metabolism in the thiotrophic clam (Solemva reidi,
444
WYLIE. CHAD R., see Kathryn L. Van Alstyne. 231
VACQUIER. VICTOR, see Youn-Ho Lee, 97
VAN ALSTYNE. KATHRYN L.. CHAD R. WYLIE. VALERIE J. PAUL, AND
KAREN MEYER, Antipredator defensed in tropical Pacific soft corals
(Coelenterata: Alcyonacea). I. Sclentes as defense against generalist
carnivorous fishes. 231
Young's modulus, 195
YUND, PHILIP O., AND MICHAEL FELDGARDEN. To thine own self be
true? Yes! Thou canst not then be false to any other. A reply to
Grosberg. 458
YUND, PHILIP O.. see Michael Feldgarden. 155
•
CONTENTS
DEVELOPMENT AND REPRODUCTION
Fong, Peter P., and John S. Pearse
Evidence for a programmed circannual life cycle
modulated by increasing daylengths in \eanthes lim-
)»«>/fl(Polychaeta:Nereidae) from central California 289
Mita, Masatoshi, and Masaru Nakamura
Ultrastructural study of an endogenous energy sub-
strate in spermatozoa of the sea urchin Hemicentrotus
pulcherrimus 298
Rivest, Brian R.
Studies on the structure and function of the larval
kidney complex of prosobranch gastropods 305
MARINE CELL BIOLOGY
Gates, Ruth I)., Garen Baghdasarian, and Leonard
Muscatine
Temperature stress causes host cell detachment in
symbiotic cnidarians: implications for coral bleach-
ing 324
NEUROBIOLOGY AND BEHAVIOR
Mercier, A. Joffre, and Rune T. Russenes
Modulation of crayfish hearts by FMRFamide-
related peptides 333
Kulkarni, Gunderao K., and Milton Fingerman
Quantitative analysis by reverse phase high perfor-
mance liquid chromatography of 5-hydroxytrypt-
amine in the central nervous system of the red
swamp crayfish, Procambarus clarkii 341
Page, Louise R.
New interpretation of a nudibranch central nervous
system based on ultrastructural analysis of neuro-
development in Melibc teimina. I. Cerebral and vis-
ceral loop ganglia 348
Page, Louise R.
New interpretation of a nudibranch central nervous
system based on ultrastructural analysis of neuro-
development in Melibe leoiiina. II. Pedal, pleural, and
labial ganglia 366
PHYSIOLOGY
Bergles, Dwight, and Sidney Tamm
Control of cilia in the branchial basket of Ciona nt-
testinalis (Ascidacea) 382
Latz, Michael I., and James F. Case
Slow photic and chemical induction of biolumines-
cence in the midwater shrimp, Sergestes similis Han-
sen 391
Fitt, W. K., and S. L. Coon
Evidence for ammonia as a natural cue for recruit-
ment of oyster larvae to oyster beds in a Georgia
salt marsh 401
Burton, Ronald S.
Proline synthesis during osmotic stress in megalopa
stage larvae of the blue crab, Callinectes sapidus . . 409
Combs, Christian A., Nicole Alford, Angela Boynton,
Mark Dvornak, and Raymond P. Henry
Behavioral regulation of hemolymph osmolarity
through selective drinking in land crabs, Birgus latro
and Gecarcoidea lalandii 416
Filers, Olaf, and Malcolm Telford
Causes and consequences of fluctuating coelomic
pressure in sea urchins 424
K m us, David W., Jeannette E. Doeller, and Jonathan
B. Wittenberg ,
Hydrogen sulfide reduction of symbiont cytochrome
<'f.s2 in gills ofSolemya reidi (Mollusca) 435
Wilmot, David B., and Russell D. Vetter
Oxygen- and nitrogen-dependent sulfur metabolism
in the thiotrophic clam Solem\a reidi 444
VIEWS AND DISCUSSION
Grosberg, Richard K.
To thine own self be true? An addendum to Feld-
garden and Yund's report on fusion and the evo-
lution of allorecognition in colonial marine inver-
tebrates 454
Yund, Philip ( ).. and Michael Feldgarden
To thine own self be true? Yes! Thou canst not then
be false to any other. A reply to Grosberg 458
Index to Volume 182 . 460
Live Music
Archive
Librivox
Free Audio
Metropolitan Museum
Cleveland
Museum of Art
Internet
Arcade
Console Living Room
Open Library
American
Libraries
TV News
Understanding
9/11